modelId
string | author
string | last_modified
timestamp[us, tz=UTC] | downloads
int64 | likes
int64 | library_name
string | tags
sequence | pipeline_tag
string | createdAt
timestamp[us, tz=UTC] | card
string |
|---|---|---|---|---|---|---|---|---|---|
QuimBallester/question_focus_model | QuimBallester | 2025-05-29T14:24:53Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"roberta",
"text-classification",
"generated_from_trainer",
"base_model:PlanTL-GOB-ES/roberta-base-bne",
"base_model:finetune:PlanTL-GOB-ES/roberta-base-bne",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | text-classification | 2025-05-29T11:01:19Z | ---
library_name: transformers
license: apache-2.0
base_model: PlanTL-GOB-ES/roberta-base-bne
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: question_focus_model
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# question_focus_model
This model is a fine-tuned version of [PlanTL-GOB-ES/roberta-base-bne](https://huggingface.co/PlanTL-GOB-ES/roberta-base-bne) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5105
- Accuracy: 0.8
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.7071 | 1.0 | 13 | 0.6947 | 0.52 |
| 0.7218 | 2.0 | 26 | 0.6797 | 0.64 |
| 0.6588 | 3.0 | 39 | 0.5430 | 0.72 |
| 0.1865 | 4.0 | 52 | 0.5481 | 0.8 |
| 0.0525 | 5.0 | 65 | 0.5105 | 0.8 |
### Framework versions
- Transformers 4.52.3
- Pytorch 2.7.0
- Datasets 3.6.0
- Tokenizers 0.21.1
|
Thalesian/Pretrain-AKK-60m | Thalesian | 2025-05-29T14:23:03Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"t5",
"text2text-generation",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text2text-generation | 2025-05-29T14:22:05Z | ---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
jinachris/PURE-PRM-7B | jinachris | 2025-05-29T14:21:42Z | 49 | 4 | null | [
"safetensors",
"qwen2",
"token-classification",
"dataset:HuggingFaceH4/prm800k-trl-dedup",
"arxiv:2504.15275",
"base_model:Qwen/Qwen2.5-Math-7B",
"base_model:finetune:Qwen/Qwen2.5-Math-7B",
"license:apache-2.0",
"region:us"
] | token-classification | 2025-02-09T07:10:09Z | ---
license: apache-2.0
base_model:
- Qwen/Qwen2.5-Math-7B
pipeline_tag: token-classification
datasets:
- HuggingFaceH4/prm800k-trl-dedup
---
> [!Warning]
> <div align="center">
> <b>
> 🚨 This repo differs from <a href=https://huggingface.co/Qwen/Qwen2.5-Math-7B-PRM800K>Qwen's PRM</a>. We trained our PRM based on <a href=https://huggingface.co/Qwen/Qwen2.5-Math-7B>Qwen2.5-Math-7B</a>, while Qwen's PRM is based on <a href=https://huggingface.co/Qwen/Qwen2.5-Math-7B-Instruct>Qwen2.5-Math-7B-Instruct</a>.
> </b>
> </div>
# PURE's PRM based on Qwen2.5-Math-7B
## Introduction
**Our PRM is used to fine-tune LLM for better math reasoning capability.** See our [PURE GitHub repo](https://github.com/CJReinforce/PURE) for more details. It is obtained by fine-tuning **Qwen2.5-Math-7B** on the training set of open-source dataset [PRM800K](https://github.com/openai/prm800k). **We choose Qwen2.5-Math-7B instead of Qwen2.5-Math-7B-Instruct to keep the base model consistent with our baselines.** We treat the original 1 and 0 labels in PRM800K as our positive labels, while -1 as negative ones. To eliminate test data contamination, we also remove the PRM800K training samples that have the same math queries in MATH test set.
## Requirements
* `transformers>=4.40.0` for Qwen2.5-Math models. The latest version is recommended.
## Quick Start
> [!Important]
>
> **PURE's PRM** is a process reward model typically used for offering feedback on the quality of reasoning and intermediate steps rather than generation.
### Prerequisites
- Step Separation: We recommend using double line breaks ("\n\n") to separate individual steps within the solution.
- Reward Computation: After each step, we insert a token "`\n`". For reward calculation, we extract the probability score of this token and subtract negative probabilities from positive probabilities, resulting in a reward value between -1 and 1. We regard steps with reward > 0 as correct, otherwise as incorrect.
### 🤗 Hugging Face Transformers
1. Here we show a code snippet to show you how to use our PRM with `transformers`:
```python
import torch
from transformers import AutoModelForTokenClassification, AutoTokenizer
def make_step_rewards(logits, token_masks):
all_scores_res = []
for sample, token_mask in zip(logits, token_masks):
# sample: (seq_len, num_labels)
probs = sample[token_mask].softmax(dim=-1) # (num_steps, 2)
process_reward = probs[:, 1] - probs[:, 0] # (num_steps,)
# weighted sum to approx. min, highly recommend when BoN eval and Fine-tuning LLM
# weight = torch.softmax(
# -process_reward / 0.1,
# dim=-1,
# )
# process_reward = weight * process_reward
all_scores_res.append(process_reward.cpu().tolist())
return all_scores_res
model_name = "jinachris/PURE-PRM-7B"
device = "auto"
tokenizer = AutoTokenizer.from_pretrained(
model_name,
trust_remote_code=True,
)
model = AutoModelForTokenClassification.from_pretrained(
model_name,
device_map=device,
torch_dtype=torch.bfloat16,
trust_remote_code=True,
).eval()
question = "Sue lives in a fun neighborhood. One weekend, the neighbors decided to play a prank on Sue. On Friday morning, the neighbors placed 18 pink plastic flamingos out on Sue's front yard. On Saturday morning, the neighbors took back one third of the flamingos, painted them white, and put these newly painted white flamingos back out on Sue's front yard. Then, on Sunday morning, they added another 18 pink plastic flamingos to the collection. At noon on Sunday, how many more pink plastic flamingos were out than white plastic flamingos?"
steps = [
"To find out how many more pink plastic flamingos were out than white plastic flamingos at noon on Sunday, we can break down the problem into steps. First, on Friday, the neighbors start with 18 pink plastic flamingos.",
"On Saturday, they take back one third of the flamingos. Since there were 18 flamingos, (1/3 \\times 18 = 6) flamingos are taken back. So, they have (18 - 6 = 12) flamingos left in their possession. Then, they paint these 6 flamingos white and put them back out on Sue's front yard. Now, Sue has the original 12 pink flamingos plus the 6 new white ones. Thus, by the end of Saturday, Sue has (12 + 6 = 18) pink flamingos and 6 white flamingos.",
"On Sunday, the neighbors add another 18 pink plastic flamingos to Sue's front yard. By the end of Sunday morning, Sue has (18 + 18 = 36) pink flamingos and still 6 white flamingos.",
"To find the difference, subtract the number of white flamingos from the number of pink flamingos: (36 - 6 = 30). Therefore, at noon on Sunday, there were 30 more pink plastic flamingos out than white plastic flamingos. The answer is (\\boxed{30})."
]
step_separator = "\n"
step_separator_token = tokenizer(
step_separator,
add_special_tokens=False,
return_tensors='pt',
)['input_ids']
input_ids = tokenizer(
question,
add_special_tokens=False,
return_tensors='pt',
)['input_ids']
score_ids = []
for step in steps:
step_ids = tokenizer(
step,
add_special_tokens=False,
return_tensors='pt',
)['input_ids']
input_ids = torch.cat(
[input_ids, step_ids, step_separator_token],
dim=-1,
)
score_ids.append(input_ids.size(-1) - 1)
input_ids = input_ids.to(model.device)
token_masks = torch.zeros_like(input_ids, dtype=torch.bool)
token_masks[0, score_ids] = True
assert torch.all(input_ids[token_masks].to("cpu") == step_separator_token)
logits = model(input_ids).logits
step_reward = make_step_rewards(logits, token_masks)
print(step_reward) # [[0.796875, 0.185546875, -0.0625, 0.078125]]
# For BoN eval,
# uncomment the weighted sum part in `make_step_rewards` func,
# then sum the rewards to get the final score (outcome reward):
# torch.tensor(step_reward).sum(dim=-1)
```
2. For evaluation using Best-of-N method or on ProcessBench and PRMBench, refer to [our github repository](https://github.com/CJReinforce/PURE/tree/verl/PRM/eval).
## Citation
If you find our work useful, we would appreciate it if you could cite our work:
```
@article{cheng2025stop,
title={Stop Summation: Min-Form Credit Assignment Is All Process Reward Model Needs for Reasoning},
author={Cheng, Jie and Qiao, Ruixi and Li, Lijun and Guo, Chao and Wang, Junle and Xiong, Gang and Lv, Yisheng and Wang, Fei-Yue},
journal={arXiv preprint arXiv:2504.15275},
year={2025}
}
``` |
bullerwins/DeepSeek-R1-0528-Qwen3-8B-GGUF | bullerwins | 2025-05-29T14:21:26Z | 0 | 1 | transformers | [
"transformers",
"gguf",
"arxiv:2501.12948",
"base_model:deepseek-ai/DeepSeek-R1-0528-Qwen3-8B",
"base_model:quantized:deepseek-ai/DeepSeek-R1-0528-Qwen3-8B",
"license:mit",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-05-29T14:19:04Z | ---
license: mit
library_name: transformers
base_model:
- deepseek-ai/DeepSeek-R1-0528-Qwen3-8B
---
# DeepSeek-R1-0528
<!-- markdownlint-disable first-line-h1 -->
<!-- markdownlint-disable html -->
<!-- markdownlint-disable no-duplicate-header -->
<div align="center">
<img src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg?raw=true" width="60%" alt="DeepSeek-V3" />
</div>
<hr>
<div align="center" style="line-height: 1;">
<a href="https://www.deepseek.com/" target="_blank" style="margin: 2px;">
<img alt="Homepage" src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/badge.svg?raw=true" style="display: inline-block; vertical-align: middle;"/>
</a>
<a href="https://chat.deepseek.com/" target="_blank" style="margin: 2px;">
<img alt="Chat" src="https://img.shields.io/badge/🤖%20Chat-DeepSeek%20R1-536af5?color=536af5&logoColor=white" style="display: inline-block; vertical-align: middle;"/>
</a>
<a href="https://huggingface.co/deepseek-ai" target="_blank" style="margin: 2px;">
<img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-DeepSeek%20AI-ffc107?color=ffc107&logoColor=white" style="display: inline-block; vertical-align: middle;"/>
</a>
</div>
<div align="center" style="line-height: 1;">
<a href="https://discord.gg/Tc7c45Zzu5" target="_blank" style="margin: 2px;">
<img alt="Discord" src="https://img.shields.io/badge/Discord-DeepSeek%20AI-7289da?logo=discord&logoColor=white&color=7289da" style="display: inline-block; vertical-align: middle;"/>
</a>
<a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/qr.jpeg?raw=true" target="_blank" style="margin: 2px;">
<img alt="Wechat" src="https://img.shields.io/badge/WeChat-DeepSeek%20AI-brightgreen?logo=wechat&logoColor=white" style="display: inline-block; vertical-align: middle;"/>
</a>
<a href="https://twitter.com/deepseek_ai" target="_blank" style="margin: 2px;">
<img alt="Twitter Follow" src="https://img.shields.io/badge/Twitter-deepseek_ai-white?logo=x&logoColor=white" style="display: inline-block; vertical-align: middle;"/>
</a>
</div>
<div align="center" style="line-height: 1;">
<a href="LICENSE" style="margin: 2px;">
<img alt="License" src="https://img.shields.io/badge/License-MIT-f5de53?&color=f5de53" style="display: inline-block; vertical-align: middle;"/>
</a>
</div>
<p align="center">
<a href="https://arxiv.org/pdf/2501.12948"><b>Paper Link</b>👁️</a>
</p>
## 1. Introduction
The DeepSeek R1 model has undergone a minor version upgrade, with the current version being DeepSeek-R1-0528. In the latest update, DeepSeek R1 has significantly improved its depth of reasoning and inference capabilities by leveraging increased computational resources and introducing algorithmic optimization mechanisms during post-training. The model has demonstrated outstanding performance across various benchmark evaluations, including mathematics, programming, and general logic. Its overall performance is now approaching that of leading models, such as O3 and Gemini 2.5 Pro.
<p align="center">
<img width="80%" src="figures/benchmark.png">
</p>
Compared to the previous version, the upgraded model shows significant improvements in handling complex reasoning tasks. For instance, in the AIME 2025 test, the model’s accuracy has increased from 70% in the previous version to 87.5% in the current version. This advancement stems from enhanced thinking depth during the reasoning process: in the AIME test set, the previous model used an average of 12K tokens per question, whereas the new version averages 23K tokens per question.
Beyond its improved reasoning capabilities, this version also offers a reduced hallucination rate, enhanced support for function calling, and better experience for vibe coding.
## 2. Evaluation Results
### DeepSeek-R1-0528
For all our models, the maximum generation length is set to 64K tokens. For benchmarks requiring sampling, we use a temperature of $0.6$, a top-p value of $0.95$, and generate 16 responses per query to estimate pass@1.
<div align="center">
| Category | Benchmark (Metric) | DeepSeek R1 | DeepSeek R1 0528
|----------|----------------------------------|-----------------|---|
| General |
| | MMLU-Redux (EM) | 92.9 | 93.4
| | MMLU-Pro (EM) | 84.0 | 85.0
| | GPQA-Diamond (Pass@1) | 71.5 | 81.0
| | SimpleQA (Correct) | 30.1 | 27.8
| | FRAMES (Acc.) | 82.5 | 83.0
| | Humanity's Last Exam (Pass@1) | 8.5 | 17.7
| Code |
| | LiveCodeBench (2408-2505) (Pass@1) | 63.5 | 73.3
| | Codeforces-Div1 (Rating) | 1530 | 1930
| | SWE Verified (Resolved) | 49.2 | 57.6
| | Aider-Polyglot (Acc.) | 53.3 | 71.6
| Math |
| | AIME 2024 (Pass@1) | 79.8 | 91.4
| | AIME 2025 (Pass@1) | 70.0 | 87.5
| | HMMT 2025 (Pass@1) | 41.7 | 79.4 |
| | CNMO 2024 (Pass@1) | 78.8 | 86.9
| Tools |
| | BFCL_v3_MultiTurn (Acc) | - | 37.0 |
| | Tau-Bench (Pass@1) | - | 53.5(Airline)/63.9(Retail)
</div>
Note: We use Agentless framework to evaluate model performance on SWE-Verified. We only evaluate text-only prompts in HLE testsets. GPT-4.1 is employed to act user role in Tau-bench evaluation.
### DeepSeek-R1-0528-Qwen3-8B
Meanwhile, we distilled the chain-of-thought from DeepSeek-R1-0528 to post-train Qwen3 8B Base, obtaining DeepSeek-R1-0528-Qwen3-8B. This model achieves state-of-the-art (SOTA) performance among open-source models on the AIME 2024, surpassing Qwen3 8B by +10.0% and matching the performance of Qwen3-235B-thinking. We believe that the chain-of-thought from DeepSeek-R1-0528 will hold significant importance for both academic research on reasoning models and industrial development focused on small-scale models.
| | AIME 24 | AIME 25 | HMMT Feb 25 | GPQA Diamond | LiveCodeBench (2408-2505) |
|--------------------------------|---------|---------|-------------|--------------|---------------------------|
| Qwen3-235B-A22B | 85.7 | 81.5 | 62.5 | 71.1 | 66.5 |
| Qwen3-32B | 81.4 | 72.9 | - | 68.4 | - |
| Qwen3-8B | 76.0 | 67.3 | - | 62.0 | - |
| Phi-4-Reasoning-Plus-14B | 81.3 | 78.0 | 53.6 | 69.3 | - |
| Gemini-2.5-Flash-Thinking-0520 | 82.3 | 72.0 | 64.2 | 82.8 | 62.3 |
| o3-mini (medium) | 79.6 | 76.7 | 53.3 | 76.8 | 65.9 |
| DeepSeek-R1-0528-Qwen3-8B | 86.0 | 76.3 | 61.5 | 61.1 | 60.5 |
## 3. Chat Website & API Platform
You can chat with DeepSeek-R1 on DeepSeek's official website: [chat.deepseek.com](https://chat.deepseek.com/sign_in), and switch on the button "DeepThink"
We also provide OpenAI-Compatible API at DeepSeek Platform: [platform.deepseek.com](https://platform.deepseek.com/)
## 4. How to Run Locally
Please visit [DeepSeek-R1](https://github.com/deepseek-ai/DeepSeek-R1) repository for more information about running DeepSeek-R1-0528 locally.
Compared to previous versions of DeepSeek-R1, the usage recommendations for DeepSeek-R1-0528 have the following changes:
1. System prompt is supported now.
2. It is not required to add "\<think\>\n" at the beginning of the output to force the model into thinking pattern.
The model architecture of DeepSeek-R1-0528-Qwen3-8B is identical to that of Qwen3-8B, but it shares the same tokenizer configuration as DeepSeek-R1-0528. This model can be run in the same manner as Qwen3-8B, but it is essential to ensure that all configuration files are sourced from our repository rather than the original Qwen3 project.
### System Prompt
In the official DeepSeek web/app, we use the same system prompt with a specific date.
```
该助手为DeepSeek-R1，由深度求索公司创造。
今天是{current date}。
```
For example,
```
该助手为DeepSeek-R1，由深度求索公司创造。
今天是2025年5月28日，星期一。
```
### Temperature
In our web and application environments, the temperature parameter $T_{model}$ is set to 0.6.
### Prompts for File Uploading and Web Search
For file uploading, please follow the template to create prompts, where {file_name}, {file_content} and {question} are arguments.
```
file_template = \
"""[file name]: {file_name}
[file content begin]
{file_content}
[file content end]
{question}"""
```
For Web Search, {search_results}, {cur_date}, and {question} are arguments.
For Chinese query, we use the prompt:
```
search_answer_zh_template = \
'''# 以下内容是基于用户发送的消息的搜索结果:
{search_results}
在我给你的搜索结果中，每个结果都是[webpage X begin]...[webpage X end]格式的，X代表每篇文章的数字索引。请在适当的情况下在句子末尾引用上下文。请按照引用编号[citation:X]的格式在答案中对应部分引用上下文。如果一句话源自多个上下文，请列出所有相关的引用编号，例如[citation:3][citation:5]，切记不要将引用集中在最后返回引用编号，而是在答案对应部分列出。
在回答时，请注意以下几点：
- 今天是{cur_date}。
- 并非搜索结果的所有内容都与用户的问题密切相关，你需要结合问题，对搜索结果进行甄别、筛选。
- 对于列举类的问题（如列举所有航班信息），尽量将答案控制在10个要点以内，并告诉用户可以查看搜索来源、获得完整信息。优先提供信息完整、最相关的列举项；如非必要，不要主动告诉用户搜索结果未提供的内容。
- 对于创作类的问题（如写论文），请务必在正文的段落中引用对应的参考编号，例如[citation:3][citation:5]，不能只在文章末尾引用。你需要解读并概括用户的题目要求，选择合适的格式，充分利用搜索结果并抽取重要信息，生成符合用户要求、极具思想深度、富有创造力与专业性的答案。你的创作篇幅需要尽可能延长，对于每一个要点的论述要推测用户的意图，给出尽可能多角度的回答要点，且务必信息量大、论述详尽。
- 如果回答很长，请尽量结构化、分段落总结。如果需要分点作答，尽量控制在5个点以内，并合并相关的内容。
- 对于客观类的问答，如果问题的答案非常简短，可以适当补充一到两句相关信息，以丰富内容。
- 你需要根据用户要求和回答内容选择合适、美观的回答格式，确保可读性强。
- 你的回答应该综合多个相关网页来回答，不能重复引用一个网页。
- 除非用户要求，否则你回答的语言需要和用户提问的语言保持一致。
# 用户消息为：
{question}'''
```
For English query, we use the prompt:
```
search_answer_en_template = \
'''# The following contents are the search results related to the user's message:
{search_results}
In the search results I provide to you, each result is formatted as [webpage X begin]...[webpage X end], where X represents the numerical index of each article. Please cite the context at the end of the relevant sentence when appropriate. Use the citation format [citation:X] in the corresponding part of your answer. If a sentence is derived from multiple contexts, list all relevant citation numbers, such as [citation:3][citation:5]. Be sure not to cluster all citations at the end; instead, include them in the corresponding parts of the answer.
When responding, please keep the following points in mind:
- Today is {cur_date}.
- Not all content in the search results is closely related to the user's question. You need to evaluate and filter the search results based on the question.
- For listing-type questions (e.g., listing all flight information), try to limit the answer to 10 key points and inform the user that they can refer to the search sources for complete information. Prioritize providing the most complete and relevant items in the list. Avoid mentioning content not provided in the search results unless necessary.
- For creative tasks (e.g., writing an essay), ensure that references are cited within the body of the text, such as [citation:3][citation:5], rather than only at the end of the text. You need to interpret and summarize the user's requirements, choose an appropriate format, fully utilize the search results, extract key information, and generate an answer that is insightful, creative, and professional. Extend the length of your response as much as possible, addressing each point in detail and from multiple perspectives, ensuring the content is rich and thorough.
- If the response is lengthy, structure it well and summarize it in paragraphs. If a point-by-point format is needed, try to limit it to 5 points and merge related content.
- For objective Q&A, if the answer is very brief, you may add one or two related sentences to enrich the content.
- Choose an appropriate and visually appealing format for your response based on the user's requirements and the content of the answer, ensuring strong readability.
- Your answer should synthesize information from multiple relevant webpages and avoid repeatedly citing the same webpage.
- Unless the user requests otherwise, your response should be in the same language as the user's question.
# The user's message is:
{question}'''
```
## 5. License
This code repository is licensed under [MIT License](LICENSE). The use of DeepSeek-R1 models is also subject to [MIT License](LICENSE). DeepSeek-R1 series (including Base and Chat) supports commercial use and distillation.
## 6. Citation
```
@misc{deepseekai2025deepseekr1incentivizingreasoningcapability,
title={DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning},
author={DeepSeek-AI},
year={2025},
eprint={2501.12948},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2501.12948},
}
```
## 7. Contact
If you have any questions, please raise an issue or contact us at [[email protected]]([email protected]). |
NikolayKozloff/DeepSeek-R1-0528-Qwen3-8B-Q8_0-GGUF | NikolayKozloff | 2025-05-29T14:17:03Z | 0 | 1 | transformers | [
"transformers",
"gguf",
"llama-cpp",
"gguf-my-repo",
"base_model:deepseek-ai/DeepSeek-R1-0528-Qwen3-8B",
"base_model:quantized:deepseek-ai/DeepSeek-R1-0528-Qwen3-8B",
"license:mit",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-05-29T14:16:25Z | ---
license: mit
library_name: transformers
tags:
- llama-cpp
- gguf-my-repo
base_model: deepseek-ai/DeepSeek-R1-0528-Qwen3-8B
---
# NikolayKozloff/DeepSeek-R1-0528-Qwen3-8B-Q8_0-GGUF
This model was converted to GGUF format from [`deepseek-ai/DeepSeek-R1-0528-Qwen3-8B`](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
Refer to the [original model card](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B) for more details on the model.
## Use with llama.cpp
Install llama.cpp through brew (works on Mac and Linux)
```bash
brew install llama.cpp
```
Invoke the llama.cpp server or the CLI.
### CLI:
```bash
llama-cli --hf-repo NikolayKozloff/DeepSeek-R1-0528-Qwen3-8B-Q8_0-GGUF --hf-file deepseek-r1-0528-qwen3-8b-q8_0.gguf -p "The meaning to life and the universe is"
```
### Server:
```bash
llama-server --hf-repo NikolayKozloff/DeepSeek-R1-0528-Qwen3-8B-Q8_0-GGUF --hf-file deepseek-r1-0528-qwen3-8b-q8_0.gguf -c 2048
```
Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well.
Step 1: Clone llama.cpp from GitHub.
```
git clone https://github.com/ggerganov/llama.cpp
```
Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux).
```
cd llama.cpp && LLAMA_CURL=1 make
```
Step 3: Run inference through the main binary.
```
./llama-cli --hf-repo NikolayKozloff/DeepSeek-R1-0528-Qwen3-8B-Q8_0-GGUF --hf-file deepseek-r1-0528-qwen3-8b-q8_0.gguf -p "The meaning to life and the universe is"
```
or
```
./llama-server --hf-repo NikolayKozloff/DeepSeek-R1-0528-Qwen3-8B-Q8_0-GGUF --hf-file deepseek-r1-0528-qwen3-8b-q8_0.gguf -c 2048
```
|
tmzconect/fotopessoal | tmzconect | 2025-05-29T14:12:54Z | 0 | 0 | null | [
"license:other",
"region:us"
] | null | 2025-05-29T13:09:53Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
--- |
benwong20418/cpu_offload_fp8_cast_bf16.py | benwong20418 | 2025-05-29T14:12:21Z | 0 | 0 | null | [
"region:us"
] | null | 2025-05-29T14:08:52Z | When trying to follow this guide to convert deepseek fp8 weight to bf16 (for converting to gguf):
https://huggingface.co/huihui-ai/DeepSeek-R1-bf16
I found out the fp8_cast_bf16.py requires >50GB vram to run. (It uses triton thus it requires nvidia gpu to run.)
I asked deepseek R1 website version to rewrite the code to reduce vram requirement, and this is the resulting code.
|
bb1070/ecm-doji-style-lr16-steps-1000 | bb1070 | 2025-05-29T14:10:23Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T13:56:28Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: ECMDOJI
---
# Ecm Doji Style Lr16 Steps 1000
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `ECMDOJI` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "ECMDOJI",
"lora_weights": "https://huggingface.co/bb1070/ecm-doji-style-lr16-steps-1000/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('bb1070/ecm-doji-style-lr16-steps-1000', weight_name='lora.safetensors')
image = pipeline('ECMDOJI').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 1000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/bb1070/ecm-doji-style-lr16-steps-1000/discussions) to add images that show off what you’ve made with this LoRA.
|
BootesVoid/cmb9ctkkj0b6j1b1yxfh92wcz_cmb9dxr0a0bsb1b1yd5lry486 | BootesVoid | 2025-05-29T14:09:55Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T14:09:53Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: ROXY
---
# Cmb9Ctkkj0B6J1B1Yxfh92Wcz_Cmb9Dxr0A0Bsb1B1Yd5Lry486
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `ROXY` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "ROXY",
"lora_weights": "https://huggingface.co/BootesVoid/cmb9ctkkj0b6j1b1yxfh92wcz_cmb9dxr0a0bsb1b1yd5lry486/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('BootesVoid/cmb9ctkkj0b6j1b1yxfh92wcz_cmb9dxr0a0bsb1b1yd5lry486', weight_name='lora.safetensors')
image = pipeline('ROXY').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 2000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/BootesVoid/cmb9ctkkj0b6j1b1yxfh92wcz_cmb9dxr0a0bsb1b1yd5lry486/discussions) to add images that show off what you’ve made with this LoRA.
|
mradermacher/TinyV-1.5B-Think-i1-GGUF | mradermacher | 2025-05-29T14:08:49Z | 0 | 0 | transformers | [
"transformers",
"gguf",
"llama-factory",
"full",
"generated_from_trainer",
"en",
"base_model:zhangchenxu/TinyV-1.5B-Think",
"base_model:quantized:zhangchenxu/TinyV-1.5B-Think",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"imatrix",
"conversational"
] | null | 2025-05-29T13:09:25Z | ---
base_model: zhangchenxu/TinyV-1.5B-Think
language:
- en
library_name: transformers
license: apache-2.0
quantized_by: mradermacher
tags:
- llama-factory
- full
- generated_from_trainer
---
## About
<!-- ### quantize_version: 2 -->
<!-- ### output_tensor_quantised: 1 -->
<!-- ### convert_type: hf -->
<!-- ### vocab_type: -->
<!-- ### tags: nicoboss -->
weighted/imatrix quants of https://huggingface.co/zhangchenxu/TinyV-1.5B-Think
<!-- provided-files -->
static quants are available at https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF
## Usage
If you are unsure how to use GGUF files, refer to one of [TheBloke's
READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
more details, including on how to concatenate multi-part files.
## Provided Quants
(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
| Link | Type | Size/GB | Notes |
|:-----|:-----|--------:|:------|
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ1_S.gguf) | i1-IQ1_S | 0.5 | for the desperate |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ1_M.gguf) | i1-IQ1_M | 0.6 | mostly desperate |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 0.6 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ2_XS.gguf) | i1-IQ2_XS | 0.7 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ2_S.gguf) | i1-IQ2_S | 0.7 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ2_M.gguf) | i1-IQ2_M | 0.7 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q2_K_S.gguf) | i1-Q2_K_S | 0.7 | very low quality |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 0.8 | lower quality |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q2_K.gguf) | i1-Q2_K | 0.8 | IQ3_XXS probably better |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ3_XS.gguf) | i1-IQ3_XS | 0.8 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q3_K_S.gguf) | i1-Q3_K_S | 0.9 | IQ3_XS probably better |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ3_S.gguf) | i1-IQ3_S | 0.9 | beats Q3_K* |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ3_M.gguf) | i1-IQ3_M | 0.9 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q3_K_M.gguf) | i1-Q3_K_M | 0.9 | IQ3_S probably better |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q3_K_L.gguf) | i1-Q3_K_L | 1.0 | IQ3_M probably better |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ4_XS.gguf) | i1-IQ4_XS | 1.0 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-IQ4_NL.gguf) | i1-IQ4_NL | 1.0 | prefer IQ4_XS |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q4_0.gguf) | i1-Q4_0 | 1.0 | fast, low quality |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q4_K_S.gguf) | i1-Q4_K_S | 1.0 | optimal size/speed/quality |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q4_K_M.gguf) | i1-Q4_K_M | 1.1 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q4_1.gguf) | i1-Q4_1 | 1.1 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q5_K_S.gguf) | i1-Q5_K_S | 1.2 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q5_K_M.gguf) | i1-Q5_K_M | 1.2 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF/resolve/main/TinyV-1.5B-Think.i1-Q6_K.gguf) | i1-Q6_K | 1.4 | practically like static Q6_K |
Here is a handy graph by ikawrakow comparing some lower-quality quant
types (lower is better):
And here are Artefact2's thoughts on the matter:
https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
## FAQ / Model Request
See https://huggingface.co/mradermacher/model_requests for some answers to
questions you might have and/or if you want some other model quantized.
## Thanks
I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
me use its servers and providing upgrades to my workstation to enable
this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to.
<!-- end -->
|
mradermacher/TinyV-1.5B-Think-GGUF | mradermacher | 2025-05-29T14:08:47Z | 0 | 0 | transformers | [
"transformers",
"gguf",
"llama-factory",
"full",
"generated_from_trainer",
"en",
"base_model:zhangchenxu/TinyV-1.5B-Think",
"base_model:quantized:zhangchenxu/TinyV-1.5B-Think",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-05-29T10:54:43Z | ---
base_model: zhangchenxu/TinyV-1.5B-Think
language:
- en
library_name: transformers
license: apache-2.0
quantized_by: mradermacher
tags:
- llama-factory
- full
- generated_from_trainer
---
## About
<!-- ### quantize_version: 2 -->
<!-- ### output_tensor_quantised: 1 -->
<!-- ### convert_type: hf -->
<!-- ### vocab_type: -->
<!-- ### tags: -->
static quants of https://huggingface.co/zhangchenxu/TinyV-1.5B-Think
<!-- provided-files -->
weighted/imatrix quants are available at https://huggingface.co/mradermacher/TinyV-1.5B-Think-i1-GGUF
## Usage
If you are unsure how to use GGUF files, refer to one of [TheBloke's
READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
more details, including on how to concatenate multi-part files.
## Provided Quants
(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
| Link | Type | Size/GB | Notes |
|:-----|:-----|--------:|:------|
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.Q2_K.gguf) | Q2_K | 0.8 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.Q3_K_S.gguf) | Q3_K_S | 0.9 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.Q3_K_M.gguf) | Q3_K_M | 0.9 | lower quality |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.Q3_K_L.gguf) | Q3_K_L | 1.0 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.IQ4_XS.gguf) | IQ4_XS | 1.0 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.Q4_K_S.gguf) | Q4_K_S | 1.0 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.Q4_K_M.gguf) | Q4_K_M | 1.1 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.Q5_K_S.gguf) | Q5_K_S | 1.2 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.Q5_K_M.gguf) | Q5_K_M | 1.2 | |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.Q6_K.gguf) | Q6_K | 1.4 | very good quality |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.Q8_0.gguf) | Q8_0 | 1.7 | fast, best quality |
| [GGUF](https://huggingface.co/mradermacher/TinyV-1.5B-Think-GGUF/resolve/main/TinyV-1.5B-Think.f16.gguf) | f16 | 3.2 | 16 bpw, overkill |
Here is a handy graph by ikawrakow comparing some lower-quality quant
types (lower is better):
And here are Artefact2's thoughts on the matter:
https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
## FAQ / Model Request
See https://huggingface.co/mradermacher/model_requests for some answers to
questions you might have and/or if you want some other model quantized.
## Thanks
I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
me use its servers and providing upgrades to my workstation to enable
this work in my free time.
<!-- end -->
|
unsloth/DeepSeek-R1-0528-Qwen3-8B | unsloth | 2025-05-29T14:04:19Z | 0 | 3 | transformers | [
"transformers",
"safetensors",
"qwen3",
"text-generation",
"unsloth",
"conversational",
"arxiv:2501.12948",
"base_model:deepseek-ai/DeepSeek-R1-0528-Qwen3-8B",
"base_model:finetune:deepseek-ai/DeepSeek-R1-0528-Qwen3-8B",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-29T13:58:27Z | ---
tags:
- unsloth
base_model:
- deepseek-ai/DeepSeek-R1-0528-Qwen3-8B
license: mit
library_name: transformers
---
<div>
<p style="margin-top: 0;margin-bottom: 0;">
<em><a href="https://docs.unsloth.ai/basics/unsloth-dynamic-v2.0-gguf">Unsloth Dynamic 2.0</a> achieves superior accuracy & outperforms other leading quants.</em>
</p>
<div style="display: flex; gap: 5px; align-items: center; ">
<a href="https://github.com/unslothai/unsloth/">
<img src="https://github.com/unslothai/unsloth/raw/main/images/unsloth%20new%20logo.png" width="133">
</a>
<a href="https://discord.gg/unsloth">
<img src="https://github.com/unslothai/unsloth/raw/main/images/Discord%20button.png" width="173">
</a>
<a href="https://docs.unsloth.ai/basics/qwen3-how-to-run-and-fine-tune">
<img src="https://raw.githubusercontent.com/unslothai/unsloth/refs/heads/main/images/documentation%20green%20button.png" width="143">
</a>
</div>
</div>
# DeepSeek-R1-0528
<!-- markdownlint-disable first-line-h1 -->
<!-- markdownlint-disable html -->
<!-- markdownlint-disable no-duplicate-header -->
<div align="center">
<img src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg?raw=true" width="60%" alt="DeepSeek-V3" />
</div>
<hr>
<div align="center" style="line-height: 1;">
<a href="https://www.deepseek.com/" target="_blank" style="margin: 2px;">
<img alt="Homepage" src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/badge.svg?raw=true" style="display: inline-block; vertical-align: middle;"/>
</a>
<a href="https://chat.deepseek.com/" target="_blank" style="margin: 2px;">
<img alt="Chat" src="https://img.shields.io/badge/🤖%20Chat-DeepSeek%20R1-536af5?color=536af5&logoColor=white" style="display: inline-block; vertical-align: middle;"/>
</a>
<a href="https://huggingface.co/deepseek-ai" target="_blank" style="margin: 2px;">
<img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-DeepSeek%20AI-ffc107?color=ffc107&logoColor=white" style="display: inline-block; vertical-align: middle;"/>
</a>
</div>
<div align="center" style="line-height: 1;">
<a href="https://discord.gg/Tc7c45Zzu5" target="_blank" style="margin: 2px;">
<img alt="Discord" src="https://img.shields.io/badge/Discord-DeepSeek%20AI-7289da?logo=discord&logoColor=white&color=7289da" style="display: inline-block; vertical-align: middle;"/>
</a>
<a href="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/qr.jpeg?raw=true" target="_blank" style="margin: 2px;">
<img alt="Wechat" src="https://img.shields.io/badge/WeChat-DeepSeek%20AI-brightgreen?logo=wechat&logoColor=white" style="display: inline-block; vertical-align: middle;"/>
</a>
<a href="https://twitter.com/deepseek_ai" target="_blank" style="margin: 2px;">
<img alt="Twitter Follow" src="https://img.shields.io/badge/Twitter-deepseek_ai-white?logo=x&logoColor=white" style="display: inline-block; vertical-align: middle;"/>
</a>
</div>
<div align="center" style="line-height: 1;">
<a href="LICENSE" style="margin: 2px;">
<img alt="License" src="https://img.shields.io/badge/License-MIT-f5de53?&color=f5de53" style="display: inline-block; vertical-align: middle;"/>
</a>
</div>
<p align="center">
<a href="https://arxiv.org/pdf/2501.12948"><b>Paper Link</b>👁️</a>
</p>
## 1. Introduction
The DeepSeek R1 model has undergone a minor version upgrade, with the current version being DeepSeek-R1-0528. In the latest update, DeepSeek R1 has significantly improved its depth of reasoning and inference capabilities by leveraging increased computational resources and introducing algorithmic optimization mechanisms during post-training. The model has demonstrated outstanding performance across various benchmark evaluations, including mathematics, programming, and general logic. Its overall performance is now approaching that of leading models, such as O3 and Gemini 2.5 Pro.
<p align="center">
<img width="80%" src="figures/benchmark.png">
</p>
Compared to the previous version, the upgraded model shows significant improvements in handling complex reasoning tasks. For instance, in the AIME 2025 test, the model’s accuracy has increased from 70% in the previous version to 87.5% in the current version. This advancement stems from enhanced thinking depth during the reasoning process: in the AIME test set, the previous model used an average of 12K tokens per question, whereas the new version averages 23K tokens per question.
Beyond its improved reasoning capabilities, this version also offers a reduced hallucination rate, enhanced support for function calling, and better experience for vibe coding.
## 2. Evaluation Results
### DeepSeek-R1-0528
For all our models, the maximum generation length is set to 64K tokens. For benchmarks requiring sampling, we use a temperature of $0.6$, a top-p value of $0.95$, and generate 16 responses per query to estimate pass@1.
<div align="center">
| Category | Benchmark (Metric) | DeepSeek R1 | DeepSeek R1 0528
|----------|----------------------------------|-----------------|---|
| General |
| | MMLU-Redux (EM) | 92.9 | 93.4
| | MMLU-Pro (EM) | 84.0 | 85.0
| | GPQA-Diamond (Pass@1) | 71.5 | 81.0
| | SimpleQA (Correct) | 30.1 | 27.8
| | FRAMES (Acc.) | 82.5 | 83.0
| | Humanity's Last Exam (Pass@1) | 8.5 | 17.7
| Code |
| | LiveCodeBench (2408-2505) (Pass@1) | 63.5 | 73.3
| | Codeforces-Div1 (Rating) | 1530 | 1930
| | SWE Verified (Resolved) | 49.2 | 57.6
| | Aider-Polyglot (Acc.) | 53.3 | 71.6
| Math |
| | AIME 2024 (Pass@1) | 79.8 | 91.4
| | AIME 2025 (Pass@1) | 70.0 | 87.5
| | HMMT 2025 (Pass@1) | 41.7 | 79.4 |
| | CNMO 2024 (Pass@1) | 78.8 | 86.9
| Tools |
| | BFCL_v3_MultiTurn (Acc) | - | 37.0 |
| | Tau-Bench (Pass@1) | - | 53.5(Airline)/63.9(Retail)
</div>
Note: We use Agentless framework to evaluate model performance on SWE-Verified. We only evaluate text-only prompts in HLE testsets. GPT-4.1 is employed to act user role in Tau-bench evaluation.
### DeepSeek-R1-0528-Qwen3-8B
Meanwhile, we distilled the chain-of-thought from DeepSeek-R1-0528 to post-train Qwen3 8B Base, obtaining DeepSeek-R1-0528-Qwen3-8B. This model achieves state-of-the-art (SOTA) performance among open-source models on the AIME 2024, surpassing Qwen3 8B by +10.0% and matching the performance of Qwen3-235B-thinking. We believe that the chain-of-thought from DeepSeek-R1-0528 will hold significant importance for both academic research on reasoning models and industrial development focused on small-scale models.
| | AIME 24 | AIME 25 | HMMT Feb 25 | GPQA Diamond | LiveCodeBench (2408-2505) |
|--------------------------------|---------|---------|-------------|--------------|---------------------------|
| Qwen3-235B-A22B | 85.7 | 81.5 | 62.5 | 71.1 | 66.5 |
| Qwen3-32B | 81.4 | 72.9 | - | 68.4 | - |
| Qwen3-8B | 76.0 | 67.3 | - | 62.0 | - |
| Phi-4-Reasoning-Plus-14B | 81.3 | 78.0 | 53.6 | 69.3 | - |
| Gemini-2.5-Flash-Thinking-0520 | 82.3 | 72.0 | 64.2 | 82.8 | 62.3 |
| o3-mini (medium) | 79.6 | 76.7 | 53.3 | 76.8 | 65.9 |
| DeepSeek-R1-0528-Qwen3-8B | 86.0 | 76.3 | 61.5 | 61.1 | 60.5 |
## 3. Chat Website & API Platform
You can chat with DeepSeek-R1 on DeepSeek's official website: [chat.deepseek.com](https://chat.deepseek.com/sign_in), and switch on the button "DeepThink"
We also provide OpenAI-Compatible API at DeepSeek Platform: [platform.deepseek.com](https://platform.deepseek.com/)
## 4. How to Run Locally
Please visit [DeepSeek-R1](https://github.com/deepseek-ai/DeepSeek-R1) repository for more information about running DeepSeek-R1-0528 locally.
Compared to previous versions of DeepSeek-R1, the usage recommendations for DeepSeek-R1-0528 have the following changes:
1. System prompt is supported now.
2. It is not required to add "\<think\>\n" at the beginning of the output to force the model into thinking pattern.
The model architecture of DeepSeek-R1-0528-Qwen3-8B is identical to that of Qwen3-8B, but it shares the same tokenizer configuration as DeepSeek-R1-0528. This model can be run in the same manner as Qwen3-8B, but it is essential to ensure that all configuration files are sourced from our repository rather than the original Qwen3 project.
### System Prompt
In the official DeepSeek web/app, we use the same system prompt with a specific date.
```
该助手为DeepSeek-R1，由深度求索公司创造。
今天是{current date}。
```
For example,
```
该助手为DeepSeek-R1，由深度求索公司创造。
今天是2025年5月28日，星期一。
```
### Temperature
In our web and application environments, the temperature parameter $T_{model}$ is set to 0.6.
### Prompts for File Uploading and Web Search
For file uploading, please follow the template to create prompts, where {file_name}, {file_content} and {question} are arguments.
```
file_template = \
"""[file name]: {file_name}
[file content begin]
{file_content}
[file content end]
{question}"""
```
For Web Search, {search_results}, {cur_date}, and {question} are arguments.
For Chinese query, we use the prompt:
```
search_answer_zh_template = \
'''# 以下内容是基于用户发送的消息的搜索结果:
{search_results}
在我给你的搜索结果中，每个结果都是[webpage X begin]...[webpage X end]格式的，X代表每篇文章的数字索引。请在适当的情况下在句子末尾引用上下文。请按照引用编号[citation:X]的格式在答案中对应部分引用上下文。如果一句话源自多个上下文，请列出所有相关的引用编号，例如[citation:3][citation:5]，切记不要将引用集中在最后返回引用编号，而是在答案对应部分列出。
在回答时，请注意以下几点：
- 今天是{cur_date}。
- 并非搜索结果的所有内容都与用户的问题密切相关，你需要结合问题，对搜索结果进行甄别、筛选。
- 对于列举类的问题（如列举所有航班信息），尽量将答案控制在10个要点以内，并告诉用户可以查看搜索来源、获得完整信息。优先提供信息完整、最相关的列举项；如非必要，不要主动告诉用户搜索结果未提供的内容。
- 对于创作类的问题（如写论文），请务必在正文的段落中引用对应的参考编号，例如[citation:3][citation:5]，不能只在文章末尾引用。你需要解读并概括用户的题目要求，选择合适的格式，充分利用搜索结果并抽取重要信息，生成符合用户要求、极具思想深度、富有创造力与专业性的答案。你的创作篇幅需要尽可能延长，对于每一个要点的论述要推测用户的意图，给出尽可能多角度的回答要点，且务必信息量大、论述详尽。
- 如果回答很长，请尽量结构化、分段落总结。如果需要分点作答，尽量控制在5个点以内，并合并相关的内容。
- 对于客观类的问答，如果问题的答案非常简短，可以适当补充一到两句相关信息，以丰富内容。
- 你需要根据用户要求和回答内容选择合适、美观的回答格式，确保可读性强。
- 你的回答应该综合多个相关网页来回答，不能重复引用一个网页。
- 除非用户要求，否则你回答的语言需要和用户提问的语言保持一致。
# 用户消息为：
{question}'''
```
For English query, we use the prompt:
```
search_answer_en_template = \
'''# The following contents are the search results related to the user's message:
{search_results}
In the search results I provide to you, each result is formatted as [webpage X begin]...[webpage X end], where X represents the numerical index of each article. Please cite the context at the end of the relevant sentence when appropriate. Use the citation format [citation:X] in the corresponding part of your answer. If a sentence is derived from multiple contexts, list all relevant citation numbers, such as [citation:3][citation:5]. Be sure not to cluster all citations at the end; instead, include them in the corresponding parts of the answer.
When responding, please keep the following points in mind:
- Today is {cur_date}.
- Not all content in the search results is closely related to the user's question. You need to evaluate and filter the search results based on the question.
- For listing-type questions (e.g., listing all flight information), try to limit the answer to 10 key points and inform the user that they can refer to the search sources for complete information. Prioritize providing the most complete and relevant items in the list. Avoid mentioning content not provided in the search results unless necessary.
- For creative tasks (e.g., writing an essay), ensure that references are cited within the body of the text, such as [citation:3][citation:5], rather than only at the end of the text. You need to interpret and summarize the user's requirements, choose an appropriate format, fully utilize the search results, extract key information, and generate an answer that is insightful, creative, and professional. Extend the length of your response as much as possible, addressing each point in detail and from multiple perspectives, ensuring the content is rich and thorough.
- If the response is lengthy, structure it well and summarize it in paragraphs. If a point-by-point format is needed, try to limit it to 5 points and merge related content.
- For objective Q&A, if the answer is very brief, you may add one or two related sentences to enrich the content.
- Choose an appropriate and visually appealing format for your response based on the user's requirements and the content of the answer, ensuring strong readability.
- Your answer should synthesize information from multiple relevant webpages and avoid repeatedly citing the same webpage.
- Unless the user requests otherwise, your response should be in the same language as the user's question.
# The user's message is:
{question}'''
```
## 5. License
This code repository is licensed under [MIT License](LICENSE). The use of DeepSeek-R1 models is also subject to [MIT License](LICENSE). DeepSeek-R1 series (including Base and Chat) supports commercial use and distillation.
## 6. Citation
```
@misc{deepseekai2025deepseekr1incentivizingreasoningcapability,
title={DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning},
author={DeepSeek-AI},
year={2025},
eprint={2501.12948},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2501.12948},
}
```
## 7. Contact
If you have any questions, please raise an issue or contact us at [[email protected]]([email protected]).
|
YuchenLi01/genParaMoreUniqueResNoGT_Qwen2.5-1.5BInstruct_dpo_ebs32_lr3e-06_beta0.4_42 | YuchenLi01 | 2025-05-29T14:01:26Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"qwen2",
"text-generation",
"alignment-handbook",
"trl",
"dpo",
"generated_from_trainer",
"conversational",
"dataset:YuchenLi01/MATH_Qwen2.5-1.5BInstruct_DPO_generatedAndParaphrasedMoreUniqueResponseNoGT",
"base_model:Qwen/Qwen2.5-1.5B-Instruct",
"base_model:finetune:Qwen/Qwen2.5-1.5B-Instruct",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-29T07:15:52Z | ---
library_name: transformers
license: apache-2.0
base_model: Qwen/Qwen2.5-1.5B-Instruct
tags:
- alignment-handbook
- trl
- dpo
- generated_from_trainer
- trl
- dpo
- generated_from_trainer
datasets:
- YuchenLi01/MATH_Qwen2.5-1.5BInstruct_DPO_generatedAndParaphrasedMoreUniqueResponseNoGT
model-index:
- name: genParaMoreUniqueResNoGT_Qwen2.5-1.5BInstruct_dpo_ebs32_lr3e-06_beta0.4_42
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# genParaMoreUniqueResNoGT_Qwen2.5-1.5BInstruct_dpo_ebs32_lr3e-06_beta0.4_42
This model is a fine-tuned version of [Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct) on the YuchenLi01/MATH_Qwen2.5-1.5BInstruct_DPO_generatedAndParaphrasedMoreUniqueResponseNoGT dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9923
- Rewards/chosen: -11.8039
- Rewards/rejected: -15.7485
- Rewards/accuracies: 0.7378
- Rewards/margins: 3.9446
- Logps/rejected: -87.1288
- Logps/chosen: -71.9740
- Logits/rejected: -1.6381
- Logits/chosen: -1.7573
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-06
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- total_train_batch_size: 32
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 2.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
| 0.7098 | 0.0135 | 20 | 0.6975 | 0.0116 | -0.0155 | 0.5488 | 0.0271 | -47.7963 | -42.4354 | -2.2052 | -2.3106 |
| 0.7363 | 0.0270 | 40 | 0.6957 | -0.0418 | -0.0641 | 0.5488 | 0.0224 | -47.9179 | -42.5687 | -2.1923 | -2.2979 |
| 0.6945 | 0.0405 | 60 | 0.6832 | -0.2329 | -0.2673 | 0.5701 | 0.0343 | -48.4257 | -43.0467 | -2.1560 | -2.2623 |
| 0.6497 | 0.0540 | 80 | 0.6708 | -0.4499 | -0.5128 | 0.5549 | 0.0629 | -49.0395 | -43.5891 | -2.1209 | -2.2281 |
| 0.5938 | 0.0675 | 100 | 0.6574 | -0.7451 | -0.8892 | 0.5915 | 0.1442 | -49.9806 | -44.3270 | -2.0633 | -2.1724 |
| 0.5491 | 0.0810 | 120 | 0.6381 | -0.7523 | -0.9836 | 0.6220 | 0.2313 | -50.2165 | -44.3450 | -2.0812 | -2.1921 |
| 0.5772 | 0.0945 | 140 | 0.6253 | -0.7988 | -1.1457 | 0.6433 | 0.3470 | -50.6219 | -44.4613 | -2.0902 | -2.2021 |
| 0.6032 | 0.1080 | 160 | 0.6110 | -0.7492 | -1.1422 | 0.6555 | 0.3930 | -50.6132 | -44.3374 | -2.1019 | -2.2147 |
| 0.5522 | 0.1215 | 180 | 0.6076 | -1.1169 | -1.6240 | 0.6829 | 0.5070 | -51.8175 | -45.2567 | -2.0429 | -2.1598 |
| 0.5428 | 0.1350 | 200 | 0.5981 | -1.1646 | -1.8209 | 0.7012 | 0.6563 | -52.3099 | -45.3759 | -2.0244 | -2.1450 |
| 0.5147 | 0.1484 | 220 | 0.5936 | -1.2658 | -1.9904 | 0.7043 | 0.7246 | -52.7337 | -45.6289 | -2.0306 | -2.1495 |
| 0.6537 | 0.1619 | 240 | 0.5936 | -1.6845 | -2.5781 | 0.7134 | 0.8937 | -54.2029 | -46.6755 | -1.9983 | -2.1199 |
| 0.5995 | 0.1754 | 260 | 0.6259 | -1.7760 | -2.7106 | 0.6555 | 0.9346 | -54.5341 | -46.9044 | -2.0381 | -2.1612 |
| 0.9123 | 0.1889 | 280 | 0.6163 | -2.0329 | -3.0730 | 0.6890 | 1.0402 | -55.4401 | -47.5465 | -1.9913 | -2.1137 |
| 0.5822 | 0.2024 | 300 | 0.6358 | -2.1738 | -3.2212 | 0.6677 | 1.0474 | -55.8106 | -47.8989 | -2.0553 | -2.1797 |
| 0.3054 | 0.2159 | 320 | 0.6567 | -3.7566 | -5.0248 | 0.6616 | 1.2682 | -60.3196 | -51.8559 | -1.8652 | -1.9890 |
| 0.8797 | 0.2294 | 340 | 0.7001 | -4.1329 | -5.3844 | 0.6646 | 1.2515 | -61.2185 | -52.7965 | -1.7727 | -1.9000 |
| 0.3861 | 0.2429 | 360 | 0.7144 | -3.4497 | -4.6169 | 0.6433 | 1.1672 | -59.2998 | -51.0885 | -1.9286 | -2.0505 |
| 0.4995 | 0.2564 | 380 | 0.7365 | -4.2502 | -5.6537 | 0.6646 | 1.4035 | -61.8919 | -53.0898 | -1.9485 | -2.0719 |
| 0.4827 | 0.2699 | 400 | 0.7037 | -4.5952 | -6.0919 | 0.6646 | 1.4967 | -62.9872 | -53.9523 | -1.9005 | -2.0232 |
| 0.2968 | 0.2834 | 420 | 0.7143 | -4.1504 | -5.5180 | 0.6463 | 1.3677 | -61.5526 | -52.8402 | -1.9489 | -2.0687 |
| 0.8907 | 0.2969 | 440 | 0.7369 | -4.9770 | -6.4202 | 0.6463 | 1.4432 | -63.8080 | -54.9068 | -1.7958 | -1.9106 |
| 0.6211 | 0.3104 | 460 | 0.7576 | -6.1123 | -7.7968 | 0.6829 | 1.6845 | -67.2496 | -57.7450 | -1.6576 | -1.7723 |
| 0.4764 | 0.3239 | 480 | 0.7325 | -5.1326 | -6.5880 | 0.6402 | 1.4553 | -64.2275 | -55.2959 | -1.7967 | -1.9149 |
| 0.7081 | 0.3374 | 500 | 0.7357 | -5.4809 | -7.1492 | 0.6433 | 1.6684 | -65.6307 | -56.1665 | -1.8434 | -1.9578 |
| 0.8628 | 0.3509 | 520 | 0.7601 | -6.1711 | -7.9028 | 0.6677 | 1.7317 | -67.5146 | -57.8921 | -1.7492 | -1.8639 |
| 0.8365 | 0.3644 | 540 | 0.7635 | -6.1656 | -7.8660 | 0.6799 | 1.7004 | -67.4226 | -57.8784 | -1.7878 | -1.8998 |
| 0.5506 | 0.3779 | 560 | 0.7630 | -5.6877 | -7.3765 | 0.6768 | 1.6888 | -66.1989 | -56.6836 | -1.8435 | -1.9586 |
| 0.827 | 0.3914 | 580 | 0.8143 | -6.4625 | -8.2914 | 0.6646 | 1.8289 | -68.4861 | -58.6206 | -1.8085 | -1.9240 |
| 0.3283 | 0.4049 | 600 | 0.8052 | -7.0358 | -8.8704 | 0.6890 | 1.8347 | -69.9337 | -60.0538 | -1.7818 | -1.8937 |
| 0.9988 | 0.4184 | 620 | 0.8165 | -7.4629 | -9.3675 | 0.6921 | 1.9046 | -71.1764 | -61.1216 | -1.7725 | -1.8848 |
| 0.5458 | 0.4318 | 640 | 0.8261 | -7.4267 | -9.4771 | 0.6982 | 2.0505 | -71.4504 | -61.0310 | -1.7991 | -1.9135 |
| 0.8243 | 0.4453 | 660 | 0.8219 | -7.2880 | -9.3348 | 0.7012 | 2.0468 | -71.0946 | -60.6844 | -1.7960 | -1.9106 |
| 0.6822 | 0.4588 | 680 | 0.8644 | -7.8936 | -9.8683 | 0.7134 | 1.9747 | -72.4284 | -62.1984 | -1.7631 | -1.8763 |
| 1.028 | 0.4723 | 700 | 0.9176 | -9.3270 | -11.4992 | 0.6890 | 2.1722 | -76.5055 | -65.7819 | -1.6213 | -1.7301 |
| 0.5688 | 0.4858 | 720 | 0.8747 | -8.5561 | -10.5672 | 0.7012 | 2.0111 | -74.1755 | -63.8545 | -1.6097 | -1.7174 |
| 0.714 | 0.4993 | 740 | 0.8826 | -8.9322 | -11.2433 | 0.6860 | 2.3111 | -75.8659 | -64.7949 | -1.6609 | -1.7705 |
| 0.755 | 0.5128 | 760 | 0.8393 | -8.3926 | -10.3983 | 0.6677 | 2.0057 | -73.7533 | -63.4459 | -1.6622 | -1.7689 |
| 0.3812 | 0.5263 | 780 | 0.8504 | -8.4765 | -10.4612 | 0.6677 | 1.9848 | -73.9107 | -63.6556 | -1.6883 | -1.7973 |
| 0.7908 | 0.5398 | 800 | 0.9168 | -9.0757 | -11.1065 | 0.6616 | 2.0308 | -75.5238 | -65.1536 | -1.5763 | -1.6818 |
| 0.4164 | 0.5533 | 820 | 0.8994 | -9.4200 | -11.6538 | 0.6890 | 2.2338 | -76.8921 | -66.0144 | -1.5816 | -1.6856 |
| 0.978 | 0.5668 | 840 | 0.9013 | -9.1077 | -11.3320 | 0.6585 | 2.2244 | -76.0877 | -65.2335 | -1.6432 | -1.7478 |
| 0.3356 | 0.5803 | 860 | 0.8793 | -9.0267 | -11.1855 | 0.6707 | 2.1588 | -75.7214 | -65.0312 | -1.5340 | -1.6340 |
| 0.651 | 0.5938 | 880 | 0.8691 | -8.7011 | -10.9951 | 0.6616 | 2.2940 | -75.2453 | -64.2171 | -1.5639 | -1.6652 |
| 0.2713 | 0.6073 | 900 | 0.9007 | -8.4464 | -10.9908 | 0.6829 | 2.5443 | -75.2345 | -63.5805 | -1.6517 | -1.7554 |
| 0.2546 | 0.6208 | 920 | 0.9749 | -9.5681 | -12.1975 | 0.6677 | 2.6294 | -78.2514 | -66.3845 | -1.5807 | -1.6810 |
| 0.1872 | 0.6343 | 940 | 0.9149 | -9.7189 | -12.1319 | 0.6616 | 2.4131 | -78.0874 | -66.7615 | -1.5820 | -1.6793 |
| 1.0918 | 0.6478 | 960 | 0.9188 | -10.2637 | -12.8306 | 0.6829 | 2.5669 | -79.8342 | -68.1236 | -1.5298 | -1.6259 |
| 0.8018 | 0.6613 | 980 | 0.8555 | -9.3842 | -11.6986 | 0.6951 | 2.3144 | -77.0042 | -65.9248 | -1.5571 | -1.6543 |
| 0.5597 | 0.6748 | 1000 | 0.8720 | -9.7656 | -12.2196 | 0.6982 | 2.4540 | -78.3065 | -66.8782 | -1.5521 | -1.6516 |
| 0.5755 | 0.6883 | 1020 | 0.9370 | -10.3906 | -12.7761 | 0.6768 | 2.3855 | -79.6978 | -68.4409 | -1.6141 | -1.7136 |
| 0.2482 | 0.7018 | 1040 | 0.8761 | -10.6673 | -12.9586 | 0.7012 | 2.2913 | -80.1541 | -69.1325 | -1.5875 | -1.6853 |
| 0.3408 | 0.7152 | 1060 | 0.9436 | -11.4063 | -14.0026 | 0.7012 | 2.5963 | -82.7640 | -70.9801 | -1.6032 | -1.7004 |
| 0.5507 | 0.7287 | 1080 | 0.9200 | -10.8737 | -13.1264 | 0.6768 | 2.2526 | -80.5735 | -69.6487 | -1.5506 | -1.6461 |
| 0.1719 | 0.7422 | 1100 | 0.9123 | -10.3203 | -12.5027 | 0.6555 | 2.1824 | -79.0144 | -68.2650 | -1.5808 | -1.6741 |
| 0.8994 | 0.7557 | 1120 | 0.9605 | -11.2090 | -13.6478 | 0.6677 | 2.4388 | -81.8770 | -70.4868 | -1.5609 | -1.6538 |
| 0.2617 | 0.7692 | 1140 | 0.9126 | -9.8912 | -12.3043 | 0.6921 | 2.4130 | -78.5182 | -67.1924 | -1.6239 | -1.7173 |
| 0.566 | 0.7827 | 1160 | 0.8975 | -10.2456 | -12.5994 | 0.6860 | 2.3538 | -79.2561 | -68.0784 | -1.6188 | -1.7091 |
| 0.4632 | 0.7962 | 1180 | 0.9409 | -11.1351 | -13.4659 | 0.6463 | 2.3308 | -81.4224 | -70.3021 | -1.5740 | -1.6659 |
| 1.0443 | 0.8097 | 1200 | 0.9132 | -10.8218 | -13.2224 | 0.6646 | 2.4007 | -80.8137 | -69.5188 | -1.5394 | -1.6349 |
| 0.7211 | 0.8232 | 1220 | 0.9058 | -10.1620 | -12.4621 | 0.7012 | 2.3000 | -78.9127 | -67.8694 | -1.5970 | -1.6952 |
| 0.4539 | 0.8367 | 1240 | 0.9035 | -10.9216 | -13.1691 | 0.6829 | 2.2475 | -80.6804 | -69.7684 | -1.5373 | -1.6338 |
| 0.4932 | 0.8502 | 1260 | 0.9137 | -10.9271 | -13.1650 | 0.6799 | 2.2379 | -80.6701 | -69.7820 | -1.5339 | -1.6310 |
| 0.4538 | 0.8637 | 1280 | 0.9164 | -10.5038 | -12.7393 | 0.6768 | 2.2356 | -79.6059 | -68.7237 | -1.5552 | -1.6528 |
| 0.3814 | 0.8772 | 1300 | 0.9254 | -10.2018 | -12.5024 | 0.6768 | 2.3005 | -79.0135 | -67.9689 | -1.5783 | -1.6753 |
| 0.4105 | 0.8907 | 1320 | 0.9508 | -10.1449 | -12.6454 | 0.6707 | 2.5005 | -79.3711 | -67.8266 | -1.6060 | -1.7035 |
| 0.2319 | 0.9042 | 1340 | 0.8939 | -9.6708 | -12.1779 | 0.6829 | 2.5070 | -78.2022 | -66.6413 | -1.6103 | -1.7073 |
| 0.3876 | 0.9177 | 1360 | 0.9378 | -10.4419 | -12.9199 | 0.6646 | 2.4781 | -80.0574 | -68.5690 | -1.5616 | -1.6557 |
| 0.1832 | 0.9312 | 1380 | 0.9283 | -10.3731 | -12.8531 | 0.6860 | 2.4800 | -79.8903 | -68.3971 | -1.5321 | -1.6255 |
| 0.4126 | 0.9447 | 1400 | 0.9057 | -10.4815 | -13.0628 | 0.6799 | 2.5812 | -80.4145 | -68.6682 | -1.5090 | -1.6053 |
| 0.9857 | 0.9582 | 1420 | 0.9436 | -11.0222 | -13.7067 | 0.6799 | 2.6845 | -82.0244 | -70.0199 | -1.5194 | -1.6173 |
| 0.6843 | 0.9717 | 1440 | 0.9204 | -10.7774 | -13.5143 | 0.6890 | 2.7369 | -81.5432 | -69.4077 | -1.5514 | -1.6515 |
| 0.2613 | 0.9852 | 1460 | 0.9273 | -9.9918 | -12.8030 | 0.7195 | 2.8111 | -79.7650 | -67.4439 | -1.6466 | -1.7508 |
| 0.3424 | 0.9987 | 1480 | 0.9197 | -10.1599 | -13.0496 | 0.7317 | 2.8898 | -80.3817 | -67.8640 | -1.6001 | -1.7011 |
| 0.0159 | 1.0121 | 1500 | 0.9066 | -10.9814 | -13.9689 | 0.7104 | 2.9876 | -82.6799 | -69.9178 | -1.5478 | -1.6470 |
| 0.0904 | 1.0256 | 1520 | 0.9502 | -11.8251 | -14.9332 | 0.6890 | 3.1082 | -85.0906 | -72.0270 | -1.5372 | -1.6373 |
| 0.2315 | 1.0391 | 1540 | 1.0112 | -11.8334 | -15.0245 | 0.6860 | 3.1911 | -85.3188 | -72.0478 | -1.6008 | -1.7058 |
| 0.0342 | 1.0526 | 1560 | 0.9967 | -11.1427 | -14.1633 | 0.6890 | 3.0207 | -83.1659 | -70.3210 | -1.6113 | -1.7178 |
| 0.0064 | 1.0661 | 1580 | 1.0481 | -11.4773 | -14.7219 | 0.6829 | 3.2446 | -84.5622 | -71.1575 | -1.5987 | -1.7054 |
| 0.0033 | 1.0796 | 1600 | 1.1210 | -12.2242 | -15.6388 | 0.7012 | 3.4146 | -86.8546 | -73.0249 | -1.6062 | -1.7153 |
| 0.1378 | 1.0931 | 1620 | 1.1277 | -12.4237 | -15.8631 | 0.6921 | 3.4394 | -87.4153 | -73.5235 | -1.6179 | -1.7290 |
| 0.2035 | 1.1066 | 1640 | 1.1075 | -12.8847 | -16.3152 | 0.6829 | 3.4305 | -88.5455 | -74.6761 | -1.5511 | -1.6598 |
| 0.0557 | 1.1201 | 1660 | 1.0968 | -13.2081 | -16.6482 | 0.6951 | 3.4401 | -89.3780 | -75.4845 | -1.5382 | -1.6452 |
| 0.122 | 1.1336 | 1680 | 1.1390 | -13.9819 | -17.5185 | 0.6921 | 3.5365 | -91.5537 | -77.4192 | -1.5434 | -1.6507 |
| 0.0428 | 1.1471 | 1700 | 1.1320 | -13.7777 | -17.4473 | 0.6982 | 3.6696 | -91.3759 | -76.9085 | -1.5807 | -1.6893 |
| 0.0581 | 1.1606 | 1720 | 1.0983 | -13.4786 | -17.0365 | 0.6921 | 3.5579 | -90.3487 | -76.1608 | -1.5582 | -1.6651 |
| 0.1992 | 1.1741 | 1740 | 1.0698 | -12.8339 | -16.3672 | 0.7134 | 3.5333 | -88.6756 | -74.5492 | -1.5737 | -1.6836 |
| 0.071 | 1.1876 | 1760 | 1.0480 | -12.0544 | -15.5962 | 0.7317 | 3.5419 | -86.7481 | -72.6002 | -1.6338 | -1.7458 |
| 0.0108 | 1.2011 | 1780 | 1.0268 | -11.8384 | -15.3437 | 0.7165 | 3.5053 | -86.1168 | -72.0604 | -1.6277 | -1.7393 |
| 0.0801 | 1.2146 | 1800 | 1.0598 | -12.2066 | -15.7010 | 0.7104 | 3.4944 | -87.0100 | -72.9808 | -1.6141 | -1.7276 |
| 0.0836 | 1.2281 | 1820 | 1.0594 | -12.0331 | -15.5395 | 0.7134 | 3.5063 | -86.6063 | -72.5472 | -1.6553 | -1.7724 |
| 0.0139 | 1.2416 | 1840 | 1.0408 | -12.3600 | -15.9404 | 0.7104 | 3.5804 | -87.6086 | -73.3644 | -1.6319 | -1.7466 |
| 0.0904 | 1.2551 | 1860 | 1.0267 | -12.3787 | -15.9427 | 0.7195 | 3.5639 | -87.6143 | -73.4112 | -1.5969 | -1.7107 |
| 0.026 | 1.2686 | 1880 | 1.0205 | -12.1565 | -15.7590 | 0.7287 | 3.6025 | -87.1552 | -72.8556 | -1.6318 | -1.7470 |
| 0.1758 | 1.2821 | 1900 | 1.0080 | -12.2911 | -15.8883 | 0.7409 | 3.5972 | -87.4783 | -73.1920 | -1.5983 | -1.7118 |
| 0.1073 | 1.2955 | 1920 | 0.9944 | -12.1375 | -15.6729 | 0.7378 | 3.5353 | -86.9397 | -72.8082 | -1.5929 | -1.7055 |
| 0.0058 | 1.3090 | 1940 | 0.9897 | -12.1333 | -15.7358 | 0.7256 | 3.6025 | -87.0970 | -72.7975 | -1.5996 | -1.7144 |
| 0.037 | 1.3225 | 1960 | 1.0029 | -12.1735 | -15.8088 | 0.7226 | 3.6353 | -87.2796 | -72.8982 | -1.6083 | -1.7235 |
| 0.0776 | 1.3360 | 1980 | 0.9931 | -12.1034 | -15.7661 | 0.7226 | 3.6627 | -87.1728 | -72.7229 | -1.6330 | -1.7497 |
| 0.1384 | 1.3495 | 2000 | 1.0111 | -12.6287 | -16.2924 | 0.7348 | 3.6637 | -88.4887 | -74.0362 | -1.6025 | -1.7182 |
| 0.0677 | 1.3630 | 2020 | 0.9950 | -12.4365 | -16.0824 | 0.7195 | 3.6459 | -87.9635 | -73.5556 | -1.6083 | -1.7233 |
| 0.1444 | 1.3765 | 2040 | 0.9950 | -12.5211 | -16.1195 | 0.7378 | 3.5985 | -88.0564 | -73.7670 | -1.5715 | -1.6856 |
| 0.141 | 1.3900 | 2060 | 0.9995 | -12.4347 | -16.0570 | 0.7409 | 3.6223 | -87.9001 | -73.5510 | -1.5604 | -1.6745 |
| 0.0546 | 1.4035 | 2080 | 1.0177 | -12.5394 | -16.2762 | 0.7348 | 3.7368 | -88.4481 | -73.8128 | -1.5829 | -1.6974 |
| 0.0527 | 1.4170 | 2100 | 1.0193 | -12.3944 | -16.2886 | 0.7317 | 3.8942 | -88.4790 | -73.4502 | -1.6122 | -1.7292 |
| 0.0131 | 1.4305 | 2120 | 1.0186 | -12.3113 | -16.2280 | 0.7256 | 3.9167 | -88.3275 | -73.2426 | -1.6353 | -1.7536 |
| 0.0053 | 1.4440 | 2140 | 1.0259 | -12.1242 | -15.9879 | 0.7287 | 3.8638 | -87.7274 | -72.7747 | -1.6805 | -1.8002 |
| 0.0374 | 1.4575 | 2160 | 1.0307 | -12.0642 | -15.9428 | 0.7165 | 3.8787 | -87.6147 | -72.6248 | -1.7024 | -1.8229 |
| 0.1171 | 1.4710 | 2180 | 1.0215 | -12.1510 | -16.0341 | 0.7317 | 3.8831 | -87.8427 | -72.8418 | -1.6942 | -1.8149 |
| 0.5026 | 1.4845 | 2200 | 1.0103 | -12.2493 | -16.0741 | 0.7256 | 3.8248 | -87.9428 | -73.0876 | -1.6660 | -1.7852 |
| 0.0387 | 1.4980 | 2220 | 1.0127 | -12.4050 | -16.2108 | 0.7439 | 3.8058 | -88.2845 | -73.4769 | -1.6343 | -1.7529 |
| 0.0975 | 1.5115 | 2240 | 1.0065 | -12.2549 | -16.0439 | 0.7378 | 3.7890 | -87.8673 | -73.1015 | -1.6191 | -1.7373 |
| 0.1411 | 1.5250 | 2260 | 0.9879 | -12.1216 | -15.9283 | 0.7439 | 3.8067 | -87.5783 | -72.7683 | -1.5930 | -1.7110 |
| 0.0151 | 1.5385 | 2280 | 0.9783 | -12.0521 | -15.7560 | 0.7378 | 3.7039 | -87.1475 | -72.5945 | -1.5685 | -1.6848 |
| 0.3175 | 1.5520 | 2300 | 0.9711 | -12.0960 | -15.7896 | 0.7317 | 3.6936 | -87.2316 | -72.7043 | -1.5593 | -1.6750 |
| 0.0208 | 1.5655 | 2320 | 0.9767 | -12.0860 | -15.8236 | 0.7317 | 3.7376 | -87.3165 | -72.6793 | -1.5798 | -1.6965 |
| 0.3457 | 1.5789 | 2340 | 0.9810 | -12.0582 | -15.8484 | 0.7348 | 3.7902 | -87.3786 | -72.6098 | -1.5878 | -1.7058 |
| 0.0218 | 1.5924 | 2360 | 0.9762 | -11.9217 | -15.6974 | 0.7348 | 3.7757 | -87.0011 | -72.2685 | -1.5899 | -1.7076 |
| 0.0608 | 1.6059 | 2380 | 0.9714 | -11.7987 | -15.6008 | 0.7378 | 3.8021 | -86.7595 | -71.9610 | -1.6042 | -1.7225 |
| 0.0403 | 1.6194 | 2400 | 0.9840 | -11.9556 | -15.8211 | 0.7165 | 3.8655 | -87.3103 | -72.3534 | -1.6113 | -1.7303 |
| 0.0087 | 1.6329 | 2420 | 0.9856 | -11.9833 | -15.8917 | 0.7256 | 3.9083 | -87.4867 | -72.4226 | -1.6188 | -1.7377 |
| 0.0136 | 1.6464 | 2440 | 0.9859 | -12.0038 | -15.9341 | 0.7317 | 3.9303 | -87.5929 | -72.4739 | -1.6266 | -1.7461 |
| 0.0442 | 1.6599 | 2460 | 0.9883 | -11.9939 | -15.8957 | 0.7287 | 3.9018 | -87.4968 | -72.4491 | -1.6189 | -1.7381 |
| 0.0802 | 1.6734 | 2480 | 0.9901 | -12.0417 | -15.9432 | 0.7287 | 3.9016 | -87.6157 | -72.5685 | -1.6140 | -1.7332 |
| 0.2235 | 1.6869 | 2500 | 0.9885 | -12.0629 | -15.9355 | 0.7378 | 3.8726 | -87.5963 | -72.6217 | -1.6067 | -1.7250 |
| 0.0092 | 1.7004 | 2520 | 0.9912 | -12.0570 | -15.9357 | 0.7378 | 3.8787 | -87.5968 | -72.6067 | -1.6083 | -1.7264 |
| 0.196 | 1.7139 | 2540 | 0.9976 | -11.9530 | -15.8323 | 0.7317 | 3.8793 | -87.3383 | -72.3469 | -1.6304 | -1.7496 |
| 0.245 | 1.7274 | 2560 | 0.9921 | -11.8814 | -15.7651 | 0.7409 | 3.8836 | -87.1703 | -72.1679 | -1.6256 | -1.7441 |
| 0.1165 | 1.7409 | 2580 | 0.9906 | -11.7657 | -15.6467 | 0.7409 | 3.8809 | -86.8743 | -71.8787 | -1.6346 | -1.7536 |
| 0.1034 | 1.7544 | 2600 | 0.9913 | -11.7350 | -15.6016 | 0.7409 | 3.8666 | -86.7616 | -71.8018 | -1.6349 | -1.7536 |
| 0.1384 | 1.7679 | 2620 | 0.9894 | -11.7153 | -15.6048 | 0.7378 | 3.8895 | -86.7695 | -71.7526 | -1.6365 | -1.7554 |
| 0.0303 | 1.7814 | 2640 | 0.9888 | -11.7269 | -15.6329 | 0.7378 | 3.9060 | -86.8397 | -71.7815 | -1.6295 | -1.7482 |
| 0.018 | 1.7949 | 2660 | 0.9908 | -11.7559 | -15.6637 | 0.7409 | 3.9079 | -86.9169 | -71.8540 | -1.6315 | -1.7504 |
| 0.0331 | 1.8084 | 2680 | 0.9917 | -11.7786 | -15.6818 | 0.7348 | 3.9032 | -86.9621 | -71.9108 | -1.6366 | -1.7558 |
| 0.0441 | 1.8219 | 2700 | 0.9898 | -11.7755 | -15.6798 | 0.7348 | 3.9042 | -86.9570 | -71.9031 | -1.6334 | -1.7525 |
| 0.0333 | 1.8354 | 2720 | 0.9943 | -11.8093 | -15.7139 | 0.7409 | 3.9046 | -87.0423 | -71.9875 | -1.6390 | -1.7586 |
| 0.0242 | 1.8489 | 2740 | 0.9900 | -11.7825 | -15.7045 | 0.7348 | 3.9220 | -87.0188 | -71.9207 | -1.6348 | -1.7538 |
| 0.0559 | 1.8623 | 2760 | 0.9903 | -11.7933 | -15.6898 | 0.7287 | 3.8965 | -86.9822 | -71.9477 | -1.6353 | -1.7548 |
| 0.1334 | 1.8758 | 2780 | 0.9913 | -11.7803 | -15.6990 | 0.7409 | 3.9188 | -87.0051 | -71.9150 | -1.6350 | -1.7541 |
| 0.0179 | 1.8893 | 2800 | 0.9913 | -11.8103 | -15.7168 | 0.7348 | 3.9064 | -87.0495 | -71.9902 | -1.6384 | -1.7580 |
| 0.0167 | 1.9028 | 2820 | 0.9912 | -11.8089 | -15.7436 | 0.7287 | 3.9346 | -87.1164 | -71.9867 | -1.6418 | -1.7611 |
| 0.0716 | 1.9163 | 2840 | 0.9923 | -11.7869 | -15.7203 | 0.7348 | 3.9335 | -87.0584 | -71.9315 | -1.6364 | -1.7555 |
| 0.0676 | 1.9298 | 2860 | 0.9928 | -11.7733 | -15.7126 | 0.7348 | 3.9392 | -87.0390 | -71.8977 | -1.6380 | -1.7574 |
| 0.0041 | 1.9433 | 2880 | 0.9941 | -11.8130 | -15.7466 | 0.7348 | 3.9336 | -87.1241 | -71.9969 | -1.6427 | -1.7629 |
| 0.0035 | 1.9568 | 2900 | 0.9938 | -11.8122 | -15.7337 | 0.7409 | 3.9215 | -87.0919 | -71.9949 | -1.6361 | -1.7550 |
| 0.0621 | 1.9703 | 2920 | 0.9903 | -11.8063 | -15.7288 | 0.7348 | 3.9225 | -87.0795 | -71.9800 | -1.6401 | -1.7593 |
| 0.199 | 1.9838 | 2940 | 0.9954 | -11.8101 | -15.7246 | 0.7409 | 3.9145 | -87.0691 | -71.9895 | -1.6363 | -1.7553 |
| 0.0343 | 1.9973 | 2960 | 0.9920 | -11.7921 | -15.7484 | 0.7378 | 3.9563 | -87.1286 | -71.9447 | -1.6381 | -1.7573 |
### Framework versions
- Transformers 4.45.2
- Pytorch 2.5.1+cu121
- Datasets 3.5.0
- Tokenizers 0.20.3
|
prasoonmhwr/ai_detection_model | prasoonmhwr | 2025-05-29T14:00:47Z | 18 | 2 | null | [
"pytorch",
"safetensors",
"roberta",
"text-classification",
"dataset:ahmadreza13/human-vs-Ai-generated-dataset",
"license:mit",
"region:us"
] | text-classification | 2025-05-10T08:56:23Z | ---
license: mit
datasets:
- ahmadreza13/human-vs-Ai-generated-dataset
pipeline_tag: text-classification
---
# AI-Generated Content Detection Model
## Model Description
This model is designed to detect AI-generated content by analyzing text using a combination of RoBERTa embeddings, Word2Vec embeddings, and engineered linguistic features.
## Model Architecture
The model utilizes a hybrid architecture that combines:
- **RoBERTa Base**: For contextual text embeddings
- **Word2Vec Embeddings**: For additional semantic information
- **Engineered Linguistic Features**: Including sentiment analysis metrics, readability scores, and lexical diversity
The model architecture consists of:
- A pre-trained RoBERTa base model with the first 6 layers frozen
- Gradient checkpointing enabled for memory efficiency
- A fully connected network that combines RoBERTa embeddings with Word2Vec and engineered features
- Three fully connected layers (512 → 128 → 1) with ReLU activations and dropout
## Training Information
- **Dataset**: https://huggingface.co/datasets/ahmadreza13/human-vs-Ai-generated-dataset
- **Training Strategy**:
- Mixed precision training with gradient accumulation
- OneCycleLR learning rate scheduler
- Early stopping based on validation F1 score
- **Hyperparameters**:
- Learning rate: 3e-5
- Batch size: 32
- Gradient accumulation steps: 2
- Dropout rate: 0.3
- Training epochs: Up to 3 with early stopping
## Performance Metrics
| Metric | Score |
|-----------|-------|
| Precision | {f1:0.9079} |
| Recall | {f1:0.9089} |
| F1 Score | {f1:0.907} |
| ROC AUC | {roc_auc:0.908} |
## Limitations
- The model's performance may vary based on the type of AI-generated content, as different AI models produce text with different characteristics
- Performance may be reduced on highly technical or domain-specific content that wasn't well-represented in the training data
- The model may produce occasional false positives on human-written content that exhibits unusually high coherence or consistency
## Ethics & Responsible Use
This model is intended to be used as a tool for:
- Research on AI-generated content characteristics
- Content moderation and filtration where transparency about content source is important
- Educational purposes to understand differences between human and AI-written content
This model should NOT be used to:
- Make high-stakes decisions without human oversight
- Discriminate against content creators
- Falsely attribute content to AI or humans with absolute certainty
## Usage Examples
```python
# Load model and tokenizer
from transformers import RobertaTokenizer, AutoModelForSequenceClassification
import torch
import numpy as np
def predict_with_huggingface_model(text, repo_id="prasoonmhwr/ai_detection_model", device="cuda"):
"""
Predicts using a model from the Hugging Face Model Hub.
Args:
text (str): The text to predict on.
repo_id (str): The repository ID of the model on Hugging Face Hub.
device (str): "cuda" if GPU is available, "cpu" otherwise
Returns:
float: The prediction probability (between 0 and 1).
"""
# 1. Load the tokenizer
tokenizer = RobertaTokenizer.from_pretrained(repo_id)
# 2. Load the model
model = AutoModelForSequenceClassification.from_pretrained(repo_id).to(device)
model.eval() # Set the model to evaluation mode
# 3. Tokenize the input text
inputs = tokenizer(text,
add_special_tokens=True,
max_length=128,
padding='max_length',
truncation=True,
return_tensors='pt').to(device) # Move inputs to device
# 4. Make the prediction (no gradient calculation needed)
with torch.no_grad():
outputs = model(**inputs)
logits = outputs.logits
probabilities = torch.sigmoid(logits).cpu().numpy().flatten() # Get probabilities, move to CPU
return probabilities[0] # Return the probability for the positive class
if __name__ == '__main__':
# Example usage:
text_to_predict = "This is a sample text to check if it was written by a human or AI"
# text_to_predict = "This text was generated by an AI model." # uncomment to test on an AI generated text
# Set the device
device = "cuda" if torch.cuda.is_available() else "cpu"
repo_id = "prasoonmhwr/ai_detection_model"
# Make the prediction
prediction = predict_with_huggingface_model(text_to_predict, repo_id, device)
# Print the result
print(f"Text: '{text_to_predict}'")
print(f"Prediction (Probability of being AI-generated): {prediction:.4f}")
if prediction > 0.5:
print("The model predicts this text is likely AI-generated.")
else:
print("The model predicts this text is likely human-generated.")
```
## Citation
If you use this model in your research, please cite:
```
@misc{ai_detection_model,
author = {Prasoon Mahawar},
title = {AI-Generated Content Detection Model},
year = {2025},
publisher = {HuggingFace},
url = {https://huggingface.co/prasoonmhwr/ai_detection_model}
}
``` |
issf/Speaker_Specific_Models | issf | 2025-05-29T13:59:25Z | 0 | 0 | null | [
"code",
"audio-classification",
"en",
"license:apache-2.0",
"region:us"
] | audio-classification | 2025-04-21T17:03:30Z | ---
license: apache-2.0
language:
- en
pipeline_tag: audio-classification
tags:
- code
--- |
Aeabds/fine-tuned-lora | Aeabds | 2025-05-29T13:59:13Z | 0 | 0 | peft | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:tiiuae/falcon-rw-1b",
"base_model:adapter:tiiuae/falcon-rw-1b",
"license:apache-2.0",
"region:us"
] | null | 2025-05-29T13:58:21Z | ---
library_name: peft
license: apache-2.0
base_model: tiiuae/falcon-rw-1b
tags:
- generated_from_trainer
model-index:
- name: fine-tuned-lora
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# fine-tuned-lora
This model is a fine-tuned version of [tiiuae/falcon-rw-1b](https://huggingface.co/tiiuae/falcon-rw-1b) on the None dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
### Training results
### Framework versions
- PEFT 0.15.2
- Transformers 4.52.2
- Pytorch 2.6.0+cu124
- Datasets 2.14.4
- Tokenizers 0.21.1 |
Varinder2110/03de1bbb-7955-4b40-bb2e-ab3ae219e815 | Varinder2110 | 2025-05-29T13:55:38Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T13:12:20Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: TOK
---
# 03De1Bbb 7955 4B40 Bb2E Ab3Ae219E815
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `TOK` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "TOK",
"lora_weights": "https://huggingface.co/Varinder2110/03de1bbb-7955-4b40-bb2e-ab3ae219e815/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('Varinder2110/03de1bbb-7955-4b40-bb2e-ab3ae219e815', weight_name='lora.safetensors')
image = pipeline('TOK').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 6000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/Varinder2110/03de1bbb-7955-4b40-bb2e-ab3ae219e815/discussions) to add images that show off what you’ve made with this LoRA.
|
Erland/vanilla-340M-4096-model-HQQ-3bit | Erland | 2025-05-29T13:48:04Z | 12 | 0 | transformers | [
"transformers",
"pytorch",
"transformer",
"text-generation",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"hqq",
"region:us"
] | text-generation | 2025-04-24T17:11:41Z | ---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
bb1070/aseem-lr16-1000steps | bb1070 | 2025-05-29T13:40:43Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T12:58:56Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: UNST
---
# Aseem Lr16 1000Steps
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `UNST` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "UNST",
"lora_weights": "https://huggingface.co/bb1070/aseem-lr16-1000steps/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('bb1070/aseem-lr16-1000steps', weight_name='lora.safetensors')
image = pipeline('UNST').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 1000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/bb1070/aseem-lr16-1000steps/discussions) to add images that show off what you’ve made with this LoRA.
|
osma77/banking77_fr_intent_model_v2 | osma77 | 2025-05-29T13:38:52Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"xlm-roberta",
"text-classification",
"generated_from_trainer",
"base_model:joeddav/xlm-roberta-large-xnli",
"base_model:finetune:joeddav/xlm-roberta-large-xnli",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | text-classification | 2025-05-29T12:14:41Z | ---
library_name: transformers
license: mit
base_model: joeddav/xlm-roberta-large-xnli
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: banking77_fr_intent_model_v2
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# banking77_fr_intent_model_v2
This model is a fine-tuned version of [joeddav/xlm-roberta-large-xnli](https://huggingface.co/joeddav/xlm-roberta-large-xnli) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8250
- Accuracy: 0.8111
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 2.0343 | 1.0 | 622 | 1.9269 | 0.6107 |
| 1.1292 | 2.0 | 1244 | 1.0095 | 0.7825 |
| 0.6979 | 3.0 | 1866 | 0.8250 | 0.8111 |
### Framework versions
- Transformers 4.51.3
- Pytorch 2.6.0+cu124
- Datasets 3.6.0
- Tokenizers 0.21.1
|
HPLT/hplt_bert_base_2_0_swh-Latn | HPLT | 2025-05-29T13:32:25Z | 7 | 0 | null | [
"pytorch",
"BERT",
"HPLT",
"encoder",
"fill-mask",
"custom_code",
"sw",
"swh",
"dataset:HPLT/HPLT2.0_cleaned",
"arxiv:2503.10267",
"license:apache-2.0",
"region:us"
] | fill-mask | 2025-04-30T18:39:56Z | ---
language:
- sw
- swh
inference: false
tags:
- BERT
- HPLT
- encoder
license: apache-2.0
datasets:
- HPLT/HPLT2.0_cleaned
pipeline_tag: fill-mask
---
# HPLT v2.0 for Swahili
<img src="https://hplt-project.org/_next/static/media/logo-hplt.d5e16ca5.svg" width=12.5%>
This is one of the encoder-only monolingual language models trained as a second release by the [HPLT project](https://hplt-project.org/).
It is a so called masked language model. In particular, we used the modification of the classic BERT model named [LTG-BERT](https://aclanthology.org/2023.findings-eacl.146/).
We present monolingual LTG-BERT models for more than 50 languages out of 191 total in the [HPLT v2.0 dataset](https://hplt-project.org/datasets/v2.0).
All the HPLT encoder-only models use the same hyper-parameters, roughly following the BERT-base setup:
- hidden size: 768
- attention heads: 12
- layers: 12
- vocabulary size: 32768
Every model uses its own tokenizer trained on language-specific HPLT data.
[The training code](https://github.com/hplt-project/HPLT-WP4).
[The training statistics of all runs](https://api.wandb.ai/links/ltg/kduj7mjn)
## Example usage
This model currently needs a custom wrapper from `modeling_ltgbert.py`, you should therefore load the model with `trust_remote_code=True`.
```python
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("HPLT/hplt_bert_base_2_0_swh-Latn")
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_swh-Latn", trust_remote_code=True)
mask_id = tokenizer.convert_tokens_to_ids("[MASK]")
input_text = tokenizer("It's a beautiful[MASK].", return_tensors="pt")
output_p = model(**input_text)
output_text = torch.where(input_text.input_ids == mask_id, output_p.logits.argmax(-1), input_text.input_ids)
# should output: '[CLS] It's a beautiful place.[SEP]'
print(tokenizer.decode(output_text[0].tolist()))
```
The following classes are currently implemented: `AutoModel`, `AutoModelMaskedLM`, `AutoModelForSequenceClassification`, `AutoModelForTokenClassification`, `AutoModelForQuestionAnswering` and `AutoModeltForMultipleChoice`.
## Intermediate checkpoints
We are releasing 10 intermediate checkpoints for each model at intervals of every 3125 training steps in separate branches. The naming convention is `stepXXX`: for example, `step18750`.
You can load a specific model revision with `transformers` using the argument `revision`:
```python
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_swh-Latn", revision="step21875", trust_remote_code=True)
```
You can access all the revisions for the models with the following code:
```python
from huggingface_hub import list_repo_refs
out = list_repo_refs("HPLT/hplt_bert_base_2_0_swh-Latn")
print([b.name for b in out.branches])
```
## Cite us
```bibtex
@inproceedings{samuel-etal-2023-trained,
title = "Trained on 100 million words and still in shape: {BERT} meets {B}ritish {N}ational {C}orpus",
author = "Samuel, David and
Kutuzov, Andrey and
{\O}vrelid, Lilja and
Velldal, Erik",
editor = "Vlachos, Andreas and
Augenstein, Isabelle",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2023",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-eacl.146",
doi = "10.18653/v1/2023.findings-eacl.146",
pages = "1954--1974"
}
```
```bibtex
@misc{burchell2025expandedmassivemultilingualdataset,
title={An Expanded Massive Multilingual Dataset for High-Performance Language Technologies},
author={Laurie Burchell and Ona de Gibert and Nikolay Arefyev and Mikko Aulamo and Marta Bañón and Pinzhen Chen and Mariia Fedorova and Liane Guillou and Barry Haddow and Jan Hajič and Jindřich Helcl and Erik Henriksson and Mateusz Klimaszewski and Ville Komulainen and Andrey Kutuzov and Joona Kytöniemi and Veronika Laippala and Petter Mæhlum and Bhavitvya Malik and Farrokh Mehryary and Vladislav Mikhailov and Nikita Moghe and Amanda Myntti and Dayyán O'Brien and Stephan Oepen and Proyag Pal and Jousia Piha and Sampo Pyysalo and Gema Ramírez-Sánchez and David Samuel and Pavel Stepachev and Jörg Tiedemann and Dušan Variš and Tereza Vojtěchová and Jaume Zaragoza-Bernabeu},
year={2025},
eprint={2503.10267},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2503.10267},
}
``` |
jellecali8/ali-speaker-embedding-model | jellecali8 | 2025-05-29T13:32:10Z | 0 | 0 | speechbrain | [
"speechbrain",
"feature-extraction",
"so",
"base_model:speechbrain/spkrec-ecapa-voxceleb",
"base_model:finetune:speechbrain/spkrec-ecapa-voxceleb",
"license:mit",
"region:us"
] | feature-extraction | 2025-05-29T09:57:21Z | ---
license: mit
language:
- so
base_model:
- speechbrain/spkrec-ecapa-voxceleb
pipeline_tag: feature-extraction
library_name: speechbrain
---
# Create README.md file first
readme_content = """
# Ali Speaker Embedding (Model Repo)
This repository contains a trained speaker embedding vector for the Somali male speaker **Ali**, extracted using the [speechbrain/spkrec-ecapa-voxceleb](https://huggingface.co/speechbrain/spkrec-ecapa-voxceleb) model.
## Model details
- Format: PyTorch .pt
- File: Ali_speaker_embedding.pt
- Embedding size: 192-dimensional
- Language: Somali
- Gender: Male
- Audio source: 300+ clips from Ali
## Usage
```python
import torch
embedding = torch.load("Ali_speaker_embedding.pt") |
BootesVoid/cmb9dkqej0bku1b1y1vgdr2em_cmb9dyz4y0bsw1b1yh38e9f0h | BootesVoid | 2025-05-29T13:29:53Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T13:29:52Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: HEYY
---
# Cmb9Dkqej0Bku1B1Y1Vgdr2Em_Cmb9Dyz4Y0Bsw1B1Yh38E9F0H
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `HEYY` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "HEYY",
"lora_weights": "https://huggingface.co/BootesVoid/cmb9dkqej0bku1b1y1vgdr2em_cmb9dyz4y0bsw1b1yh38e9f0h/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('BootesVoid/cmb9dkqej0bku1b1y1vgdr2em_cmb9dyz4y0bsw1b1yh38e9f0h', weight_name='lora.safetensors')
image = pipeline('HEYY').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 2000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/BootesVoid/cmb9dkqej0bku1b1y1vgdr2em_cmb9dyz4y0bsw1b1yh38e9f0h/discussions) to add images that show off what you’ve made with this LoRA.
|
BootesVoid/cmb8vkcpy00kc1b1ywfqyfpky_cmb94dmcy05us1b1yjlygz4st | BootesVoid | 2025-05-29T13:28:29Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T13:28:28Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: ivory
---
# Cmb8Vkcpy00Kc1B1Ywfqyfpky_Cmb94Dmcy05Us1B1Yjlygz4St
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `ivory` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "ivory",
"lora_weights": "https://huggingface.co/BootesVoid/cmb8vkcpy00kc1b1ywfqyfpky_cmb94dmcy05us1b1yjlygz4st/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('BootesVoid/cmb8vkcpy00kc1b1ywfqyfpky_cmb94dmcy05us1b1yjlygz4st', weight_name='lora.safetensors')
image = pipeline('ivory').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 2000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/BootesVoid/cmb8vkcpy00kc1b1ywfqyfpky_cmb94dmcy05us1b1yjlygz4st/discussions) to add images that show off what you’ve made with this LoRA.
|
HPLT/hplt_bert_base_2_0_ukr-Cyrl | HPLT | 2025-05-29T13:26:14Z | 40 | 0 | null | [
"pytorch",
"BERT",
"HPLT",
"encoder",
"fill-mask",
"custom_code",
"uk",
"ukr",
"dataset:HPLT/HPLT2.0_cleaned",
"arxiv:2503.10267",
"license:apache-2.0",
"region:us"
] | fill-mask | 2025-02-22T22:30:57Z | ---
language:
- uk
- ukr
inference: false
tags:
- BERT
- HPLT
- encoder
license: apache-2.0
datasets:
- HPLT/HPLT2.0_cleaned
pipeline_tag: fill-mask
---
# HPLT v2.0 BERT for Ukrainian
<img src="https://hplt-project.org/_next/static/media/logo-hplt.d5e16ca5.svg" width=12.5%>
This is one of the encoder-only monolingual language models trained as a second release by the [HPLT project](https://hplt-project.org/).
It is a so called masked language model. In particular, we used the modification of the classic BERT model named [LTG-BERT](https://aclanthology.org/2023.findings-eacl.146/).
We present monolingual LTG-BERT models for more than 50 languages out of 191 total in the [HPLT v2.0 dataset](https://hplt-project.org/datasets/v2.0).
All the HPLT encoder-only models use the same hyper-parameters, roughly following the BERT-base setup:
- hidden size: 768
- attention heads: 12
- layers: 12
- vocabulary size: 32768
Every model uses its own tokenizer trained on language-specific HPLT data.
[The training code](https://github.com/hplt-project/HPLT-WP4).
[The training statistics of all runs](https://api.wandb.ai/links/ltg/kduj7mjn)
## Example usage (tested with `transformers==4.46.1` and `tokenizers==0.20.1`)
This model currently needs a custom wrapper from `modeling_ltgbert.py`, you should therefore load the model with `trust_remote_code=True`.
```python
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("HPLT/hplt_bert_base_2_0_ukr-Cyrl")
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_ukr-Cyrl", trust_remote_code=True)
mask_id = tokenizer.convert_tokens_to_ids("[MASK]")
input_text = tokenizer("It's a beautiful[MASK].", return_tensors="pt")
output_p = model(**input_text)
output_text = torch.where(input_text.input_ids == mask_id, output_p.logits.argmax(-1), input_text.input_ids)
# should output: '[CLS] It's a beautiful place.[SEP]'
print(tokenizer.decode(output_text[0].tolist(), clean_up_tokenization_spaces=True))
```
The following classes are currently implemented: `AutoModel`, `AutoModelMaskedLM`, `AutoModelForSequenceClassification`, `AutoModelForTokenClassification`, `AutoModelForQuestionAnswering` and `AutoModeltForMultipleChoice`.
## Intermediate checkpoints
We are releasing 10 intermediate checkpoints for each model at intervals of every 3125 training steps in separate branches. The naming convention is `stepXXX`: for example, `step18750`.
You can load a specific model revision with `transformers` using the argument `revision`:
```python
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_ukr-Cyrl", revision="step21875", trust_remote_code=True)
```
You can access all the revisions for the models with the following code:
```python
from huggingface_hub import list_repo_refs
out = list_repo_refs("HPLT/hplt_bert_base_2_0_ukr-Cyrl")
print([b.name for b in out.branches])
```
## Cite us
```bibtex
@inproceedings{samuel-etal-2023-trained,
title = "Trained on 100 million words and still in shape: {BERT} meets {B}ritish {N}ational {C}orpus",
author = "Samuel, David and
Kutuzov, Andrey and
{\O}vrelid, Lilja and
Velldal, Erik",
editor = "Vlachos, Andreas and
Augenstein, Isabelle",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2023",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-eacl.146",
doi = "10.18653/v1/2023.findings-eacl.146",
pages = "1954--1974"
}
```
```bibtex
@misc{burchell2025expandedmassivemultilingualdataset,
title={An Expanded Massive Multilingual Dataset for High-Performance Language Technologies},
author={Laurie Burchell and Ona de Gibert and Nikolay Arefyev and Mikko Aulamo and Marta Bañón and Pinzhen Chen and Mariia Fedorova and Liane Guillou and Barry Haddow and Jan Hajič and Jindřich Helcl and Erik Henriksson and Mateusz Klimaszewski and Ville Komulainen and Andrey Kutuzov and Joona Kytöniemi and Veronika Laippala and Petter Mæhlum and Bhavitvya Malik and Farrokh Mehryary and Vladislav Mikhailov and Nikita Moghe and Amanda Myntti and Dayyán O'Brien and Stephan Oepen and Proyag Pal and Jousia Piha and Sampo Pyysalo and Gema Ramírez-Sánchez and David Samuel and Pavel Stepachev and Jörg Tiedemann and Dušan Variš and Tereza Vojtěchová and Jaume Zaragoza-Bernabeu},
year={2025},
eprint={2503.10267},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2503.10267},
}
```
|
BETACAROTENE343/kaya05 | BETACAROTENE343 | 2025-05-29T13:25:51Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T13:01:27Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: kaya05
---
# Kaya05
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `kaya05` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "kaya05",
"lora_weights": "https://huggingface.co/BETACAROTENE343/kaya05/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('BETACAROTENE343/kaya05', weight_name='lora.safetensors')
image = pipeline('kaya05').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 2000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/BETACAROTENE343/kaya05/discussions) to add images that show off what you’ve made with this LoRA.
|
HPLT/hplt_bert_base_2_0_slk-Latn | HPLT | 2025-05-29T13:25:20Z | 6 | 0 | null | [
"pytorch",
"BERT",
"HPLT",
"encoder",
"fill-mask",
"custom_code",
"sk",
"slk",
"dataset:HPLT/HPLT2.0_cleaned",
"arxiv:2503.10267",
"license:apache-2.0",
"region:us"
] | fill-mask | 2025-02-22T22:55:49Z | ---
language:
- sk
- slk
inference: false
tags:
- BERT
- HPLT
- encoder
license: apache-2.0
datasets:
- HPLT/HPLT2.0_cleaned
pipeline_tag: fill-mask
---
# HPLT v2.0 BERT for Slovak
<img src="https://hplt-project.org/_next/static/media/logo-hplt.d5e16ca5.svg" width=12.5%>
This is one of the encoder-only monolingual language models trained as a second release by the [HPLT project](https://hplt-project.org/).
It is a so called masked language model. In particular, we used the modification of the classic BERT model named [LTG-BERT](https://aclanthology.org/2023.findings-eacl.146/).
We present monolingual LTG-BERT models for more than 50 languages out of 191 total in the [HPLT v2.0 dataset](https://hplt-project.org/datasets/v2.0).
All the HPLT encoder-only models use the same hyper-parameters, roughly following the BERT-base setup:
- hidden size: 768
- attention heads: 12
- layers: 12
- vocabulary size: 32768
Every model uses its own tokenizer trained on language-specific HPLT data.
[The training code](https://github.com/hplt-project/HPLT-WP4).
[The training statistics of all runs](https://api.wandb.ai/links/ltg/kduj7mjn)
## Example usage (tested with `transformers==4.46.1` and `tokenizers==0.20.1`)
This model currently needs a custom wrapper from `modeling_ltgbert.py`, you should therefore load the model with `trust_remote_code=True`.
```python
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("HPLT/hplt_bert_base_2_0_slk-Latn")
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_slk-Latn", trust_remote_code=True)
mask_id = tokenizer.convert_tokens_to_ids("[MASK]")
input_text = tokenizer("It's a beautiful[MASK].", return_tensors="pt")
output_p = model(**input_text)
output_text = torch.where(input_text.input_ids == mask_id, output_p.logits.argmax(-1), input_text.input_ids)
# should output: '[CLS] It's a beautiful place.[SEP]'
print(tokenizer.decode(output_text[0].tolist(), clean_up_tokenization_spaces=True))
```
The following classes are currently implemented: `AutoModel`, `AutoModelMaskedLM`, `AutoModelForSequenceClassification`, `AutoModelForTokenClassification`, `AutoModelForQuestionAnswering` and `AutoModeltForMultipleChoice`.
## Intermediate checkpoints
We are releasing 10 intermediate checkpoints for each model at intervals of every 3125 training steps in separate branches. The naming convention is `stepXXX`: for example, `step18750`.
You can load a specific model revision with `transformers` using the argument `revision`:
```python
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_slk-Latn", revision="step21875", trust_remote_code=True)
```
You can access all the revisions for the models with the following code:
```python
from huggingface_hub import list_repo_refs
out = list_repo_refs("HPLT/hplt_bert_base_2_0_slk-Latn")
print([b.name for b in out.branches])
```
## Cite us
```bibtex
@inproceedings{samuel-etal-2023-trained,
title = "Trained on 100 million words and still in shape: {BERT} meets {B}ritish {N}ational {C}orpus",
author = "Samuel, David and
Kutuzov, Andrey and
{\O}vrelid, Lilja and
Velldal, Erik",
editor = "Vlachos, Andreas and
Augenstein, Isabelle",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2023",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-eacl.146",
doi = "10.18653/v1/2023.findings-eacl.146",
pages = "1954--1974"
}
```
```bibtex
@misc{burchell2025expandedmassivemultilingualdataset,
title={An Expanded Massive Multilingual Dataset for High-Performance Language Technologies},
author={Laurie Burchell and Ona de Gibert and Nikolay Arefyev and Mikko Aulamo and Marta Bañón and Pinzhen Chen and Mariia Fedorova and Liane Guillou and Barry Haddow and Jan Hajič and Jindřich Helcl and Erik Henriksson and Mateusz Klimaszewski and Ville Komulainen and Andrey Kutuzov and Joona Kytöniemi and Veronika Laippala and Petter Mæhlum and Bhavitvya Malik and Farrokh Mehryary and Vladislav Mikhailov and Nikita Moghe and Amanda Myntti and Dayyán O'Brien and Stephan Oepen and Proyag Pal and Jousia Piha and Sampo Pyysalo and Gema Ramírez-Sánchez and David Samuel and Pavel Stepachev and Jörg Tiedemann and Dušan Variš and Tereza Vojtěchová and Jaume Zaragoza-Bernabeu},
year={2025},
eprint={2503.10267},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2503.10267},
}
```
|
HPLT/hplt_bert_base_2_0_rus-Cyrl | HPLT | 2025-05-29T13:25:11Z | 0 | 0 | null | [
"pytorch",
"BERT",
"HPLT",
"encoder",
"fill-mask",
"custom_code",
"ru",
"rus",
"dataset:HPLT/HPLT2.0_cleaned",
"arxiv:2503.10267",
"license:apache-2.0",
"region:us"
] | fill-mask | 2025-02-22T22:55:34Z | ---
language:
- ru
- rus
inference: false
tags:
- BERT
- HPLT
- encoder
license: apache-2.0
datasets:
- HPLT/HPLT2.0_cleaned
pipeline_tag: fill-mask
---
# HPLT v2.0 BERT for Russian
<img src="https://hplt-project.org/_next/static/media/logo-hplt.d5e16ca5.svg" width=12.5%>
This is one of the encoder-only monolingual language models trained as a second release by the [HPLT project](https://hplt-project.org/).
It is a so called masked language model. In particular, we used the modification of the classic BERT model named [LTG-BERT](https://aclanthology.org/2023.findings-eacl.146/).
We present monolingual LTG-BERT models for more than 50 languages out of 191 total in the [HPLT v2.0 dataset](https://hplt-project.org/datasets/v2.0).
All the HPLT encoder-only models use the same hyper-parameters, roughly following the BERT-base setup:
- hidden size: 768
- attention heads: 12
- layers: 12
- vocabulary size: 32768
Every model uses its own tokenizer trained on language-specific HPLT data.
[The training code](https://github.com/hplt-project/HPLT-WP4).
[The training statistics of all runs](https://api.wandb.ai/links/ltg/kduj7mjn)
## Example usage (tested with `transformers==4.46.1` and `tokenizers==0.20.1`)
This model currently needs a custom wrapper from `modeling_ltgbert.py`, you should therefore load the model with `trust_remote_code=True`.
```python
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("HPLT/hplt_bert_base_2_0_rus-Cyrl")
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_rus-Cyrl", trust_remote_code=True)
mask_id = tokenizer.convert_tokens_to_ids("[MASK]")
input_text = tokenizer("It's a beautiful[MASK].", return_tensors="pt")
output_p = model(**input_text)
output_text = torch.where(input_text.input_ids == mask_id, output_p.logits.argmax(-1), input_text.input_ids)
# should output: '[CLS] It's a beautiful place.[SEP]'
print(tokenizer.decode(output_text[0].tolist(), clean_up_tokenization_spaces=True))
```
The following classes are currently implemented: `AutoModel`, `AutoModelMaskedLM`, `AutoModelForSequenceClassification`, `AutoModelForTokenClassification`, `AutoModelForQuestionAnswering` and `AutoModeltForMultipleChoice`.
## Intermediate checkpoints
We are releasing 10 intermediate checkpoints for each model at intervals of every 3125 training steps in separate branches. The naming convention is `stepXXX`: for example, `step18750`.
You can load a specific model revision with `transformers` using the argument `revision`:
```python
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_rus-Cyrl", revision="step21875", trust_remote_code=True)
```
You can access all the revisions for the models with the following code:
```python
from huggingface_hub import list_repo_refs
out = list_repo_refs("HPLT/hplt_bert_base_2_0_rus-Cyrl")
print([b.name for b in out.branches])
```
## Cite us
```bibtex
@inproceedings{samuel-etal-2023-trained,
title = "Trained on 100 million words and still in shape: {BERT} meets {B}ritish {N}ational {C}orpus",
author = "Samuel, David and
Kutuzov, Andrey and
{\O}vrelid, Lilja and
Velldal, Erik",
editor = "Vlachos, Andreas and
Augenstein, Isabelle",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2023",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-eacl.146",
doi = "10.18653/v1/2023.findings-eacl.146",
pages = "1954--1974"
}
```
```bibtex
@misc{burchell2025expandedmassivemultilingualdataset,
title={An Expanded Massive Multilingual Dataset for High-Performance Language Technologies},
author={Laurie Burchell and Ona de Gibert and Nikolay Arefyev and Mikko Aulamo and Marta Bañón and Pinzhen Chen and Mariia Fedorova and Liane Guillou and Barry Haddow and Jan Hajič and Jindřich Helcl and Erik Henriksson and Mateusz Klimaszewski and Ville Komulainen and Andrey Kutuzov and Joona Kytöniemi and Veronika Laippala and Petter Mæhlum and Bhavitvya Malik and Farrokh Mehryary and Vladislav Mikhailov and Nikita Moghe and Amanda Myntti and Dayyán O'Brien and Stephan Oepen and Proyag Pal and Jousia Piha and Sampo Pyysalo and Gema Ramírez-Sánchez and David Samuel and Pavel Stepachev and Jörg Tiedemann and Dušan Variš and Tereza Vojtěchová and Jaume Zaragoza-Bernabeu},
year={2025},
eprint={2503.10267},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2503.10267},
}
```
|
HPLT/hplt_bert_base_2_0_lit-Latn | HPLT | 2025-05-29T13:23:28Z | 4 | 0 | null | [
"pytorch",
"BERT",
"HPLT",
"encoder",
"fill-mask",
"custom_code",
"lt",
"lit",
"dataset:HPLT/HPLT2.0_cleaned",
"arxiv:2503.10267",
"license:apache-2.0",
"region:us"
] | fill-mask | 2025-02-22T22:52:28Z | ---
language:
- lt
- lit
inference: false
tags:
- BERT
- HPLT
- encoder
license: apache-2.0
datasets:
- HPLT/HPLT2.0_cleaned
pipeline_tag: fill-mask
---
# HPLT v2.0 BERT for Lithuanian
<img src="https://hplt-project.org/_next/static/media/logo-hplt.d5e16ca5.svg" width=12.5%>
This is one of the encoder-only monolingual language models trained as a second release by the [HPLT project](https://hplt-project.org/).
It is a so called masked language model. In particular, we used the modification of the classic BERT model named [LTG-BERT](https://aclanthology.org/2023.findings-eacl.146/).
We present monolingual LTG-BERT models for more than 50 languages out of 191 total in the [HPLT v2.0 dataset](https://hplt-project.org/datasets/v2.0).
All the HPLT encoder-only models use the same hyper-parameters, roughly following the BERT-base setup:
- hidden size: 768
- attention heads: 12
- layers: 12
- vocabulary size: 32768
Every model uses its own tokenizer trained on language-specific HPLT data.
[The training code](https://github.com/hplt-project/HPLT-WP4).
[The training statistics of all runs](https://api.wandb.ai/links/ltg/kduj7mjn)
## Example usage (tested with `transformers==4.46.1` and `tokenizers==0.20.1`)
This model currently needs a custom wrapper from `modeling_ltgbert.py`, you should therefore load the model with `trust_remote_code=True`.
```python
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("HPLT/hplt_bert_base_2_0_lit-Latn")
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_lit-Latn", trust_remote_code=True)
mask_id = tokenizer.convert_tokens_to_ids("[MASK]")
input_text = tokenizer("It's a beautiful[MASK].", return_tensors="pt")
output_p = model(**input_text)
output_text = torch.where(input_text.input_ids == mask_id, output_p.logits.argmax(-1), input_text.input_ids)
# should output: '[CLS] It's a beautiful place.[SEP]'
print(tokenizer.decode(output_text[0].tolist(), clean_up_tokenization_spaces=True))
```
The following classes are currently implemented: `AutoModel`, `AutoModelMaskedLM`, `AutoModelForSequenceClassification`, `AutoModelForTokenClassification`, `AutoModelForQuestionAnswering` and `AutoModeltForMultipleChoice`.
## Intermediate checkpoints
We are releasing 10 intermediate checkpoints for each model at intervals of every 3125 training steps in separate branches. The naming convention is `stepXXX`: for example, `step18750`.
You can load a specific model revision with `transformers` using the argument `revision`:
```python
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_lit-Latn", revision="step21875", trust_remote_code=True)
```
You can access all the revisions for the models with the following code:
```python
from huggingface_hub import list_repo_refs
out = list_repo_refs("HPLT/hplt_bert_base_2_0_lit-Latn")
print([b.name for b in out.branches])
```
## Cite us
```bibtex
@inproceedings{samuel-etal-2023-trained,
title = "Trained on 100 million words and still in shape: {BERT} meets {B}ritish {N}ational {C}orpus",
author = "Samuel, David and
Kutuzov, Andrey and
{\O}vrelid, Lilja and
Velldal, Erik",
editor = "Vlachos, Andreas and
Augenstein, Isabelle",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2023",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-eacl.146",
doi = "10.18653/v1/2023.findings-eacl.146",
pages = "1954--1974"
}
```
```bibtex
@misc{burchell2025expandedmassivemultilingualdataset,
title={An Expanded Massive Multilingual Dataset for High-Performance Language Technologies},
author={Laurie Burchell and Ona de Gibert and Nikolay Arefyev and Mikko Aulamo and Marta Bañón and Pinzhen Chen and Mariia Fedorova and Liane Guillou and Barry Haddow and Jan Hajič and Jindřich Helcl and Erik Henriksson and Mateusz Klimaszewski and Ville Komulainen and Andrey Kutuzov and Joona Kytöniemi and Veronika Laippala and Petter Mæhlum and Bhavitvya Malik and Farrokh Mehryary and Vladislav Mikhailov and Nikita Moghe and Amanda Myntti and Dayyán O'Brien and Stephan Oepen and Proyag Pal and Jousia Piha and Sampo Pyysalo and Gema Ramírez-Sánchez and David Samuel and Pavel Stepachev and Jörg Tiedemann and Dušan Variš and Tereza Vojtěchová and Jaume Zaragoza-Bernabeu},
year={2025},
eprint={2503.10267},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2503.10267},
}
```
|
HPLT/hplt_bert_base_2_0_jpn-Jpan | HPLT | 2025-05-29T13:22:56Z | 3 | 0 | null | [
"pytorch",
"BERT",
"HPLT",
"encoder",
"fill-mask",
"custom_code",
"ja",
"jpn",
"dataset:HPLT/HPLT2.0_cleaned",
"arxiv:2503.10267",
"license:apache-2.0",
"region:us"
] | fill-mask | 2025-02-22T22:51:43Z | ---
language:
- ja
- jpn
inference: false
tags:
- BERT
- HPLT
- encoder
license: apache-2.0
datasets:
- HPLT/HPLT2.0_cleaned
pipeline_tag: fill-mask
---
# HPLT v2.0 BERT for Japanese
<img src="https://hplt-project.org/_next/static/media/logo-hplt.d5e16ca5.svg" width=12.5%>
This is one of the encoder-only monolingual language models trained as a second release by the [HPLT project](https://hplt-project.org/).
It is a so called masked language model. In particular, we used the modification of the classic BERT model named [LTG-BERT](https://aclanthology.org/2023.findings-eacl.146/).
We present monolingual LTG-BERT models for more than 50 languages out of 191 total in the [HPLT v2.0 dataset](https://hplt-project.org/datasets/v2.0).
All the HPLT encoder-only models use the same hyper-parameters, roughly following the BERT-base setup:
- hidden size: 768
- attention heads: 12
- layers: 12
- vocabulary size: 32768
Every model uses its own tokenizer trained on language-specific HPLT data.
[The training code](https://github.com/hplt-project/HPLT-WP4).
[The training statistics of all runs](https://api.wandb.ai/links/ltg/kduj7mjn)
## Example usage (tested with `transformers==4.46.1` and `tokenizers==0.20.1`)
This model currently needs a custom wrapper from `modeling_ltgbert.py`, you should therefore load the model with `trust_remote_code=True`.
```python
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("HPLT/hplt_bert_base_2_0_jpn-Jpan")
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_jpn-Jpan", trust_remote_code=True)
mask_id = tokenizer.convert_tokens_to_ids("[MASK]")
input_text = tokenizer("It's a beautiful[MASK].", return_tensors="pt")
output_p = model(**input_text)
output_text = torch.where(input_text.input_ids == mask_id, output_p.logits.argmax(-1), input_text.input_ids)
# should output: '[CLS] It's a beautiful place.[SEP]'
print(tokenizer.decode(output_text[0].tolist(), clean_up_tokenization_spaces=True))
```
The following classes are currently implemented: `AutoModel`, `AutoModelMaskedLM`, `AutoModelForSequenceClassification`, `AutoModelForTokenClassification`, `AutoModelForQuestionAnswering` and `AutoModeltForMultipleChoice`.
## Intermediate checkpoints
We are releasing 10 intermediate checkpoints for each model at intervals of every 3125 training steps in separate branches. The naming convention is `stepXXX`: for example, `step18750`.
You can load a specific model revision with `transformers` using the argument `revision`:
```python
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_jpn-Jpan", revision="step21875", trust_remote_code=True)
```
You can access all the revisions for the models with the following code:
```python
from huggingface_hub import list_repo_refs
out = list_repo_refs("HPLT/hplt_bert_base_2_0_jpn-Jpan")
print([b.name for b in out.branches])
```
## Cite us
```bibtex
@inproceedings{samuel-etal-2023-trained,
title = "Trained on 100 million words and still in shape: {BERT} meets {B}ritish {N}ational {C}orpus",
author = "Samuel, David and
Kutuzov, Andrey and
{\O}vrelid, Lilja and
Velldal, Erik",
editor = "Vlachos, Andreas and
Augenstein, Isabelle",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2023",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-eacl.146",
doi = "10.18653/v1/2023.findings-eacl.146",
pages = "1954--1974"
}
```
```bibtex
@misc{burchell2025expandedmassivemultilingualdataset,
title={An Expanded Massive Multilingual Dataset for High-Performance Language Technologies},
author={Laurie Burchell and Ona de Gibert and Nikolay Arefyev and Mikko Aulamo and Marta Bañón and Pinzhen Chen and Mariia Fedorova and Liane Guillou and Barry Haddow and Jan Hajič and Jindřich Helcl and Erik Henriksson and Mateusz Klimaszewski and Ville Komulainen and Andrey Kutuzov and Joona Kytöniemi and Veronika Laippala and Petter Mæhlum and Bhavitvya Malik and Farrokh Mehryary and Vladislav Mikhailov and Nikita Moghe and Amanda Myntti and Dayyán O'Brien and Stephan Oepen and Proyag Pal and Jousia Piha and Sampo Pyysalo and Gema Ramírez-Sánchez and David Samuel and Pavel Stepachev and Jörg Tiedemann and Dušan Variš and Tereza Vojtěchová and Jaume Zaragoza-Bernabeu},
year={2025},
eprint={2503.10267},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2503.10267},
}
```
|
HPLT/hplt_bert_base_2_0_ita-Latn | HPLT | 2025-05-29T13:22:48Z | 3 | 0 | null | [
"pytorch",
"BERT",
"HPLT",
"encoder",
"fill-mask",
"custom_code",
"it",
"ita",
"dataset:HPLT/HPLT2.0_cleaned",
"arxiv:2503.10267",
"license:apache-2.0",
"region:us"
] | fill-mask | 2025-02-22T22:51:28Z | ---
language:
- it
- ita
inference: false
tags:
- BERT
- HPLT
- encoder
license: apache-2.0
datasets:
- HPLT/HPLT2.0_cleaned
pipeline_tag: fill-mask
---
# HPLT v2.0 BERT for Italian
<img src="https://hplt-project.org/_next/static/media/logo-hplt.d5e16ca5.svg" width=12.5%>
This is one of the encoder-only monolingual language models trained as a second release by the [HPLT project](https://hplt-project.org/).
It is a so called masked language model. In particular, we used the modification of the classic BERT model named [LTG-BERT](https://aclanthology.org/2023.findings-eacl.146/).
We present monolingual LTG-BERT models for more than 50 languages out of 191 total in the [HPLT v2.0 dataset](https://hplt-project.org/datasets/v2.0).
All the HPLT encoder-only models use the same hyper-parameters, roughly following the BERT-base setup:
- hidden size: 768
- attention heads: 12
- layers: 12
- vocabulary size: 32768
Every model uses its own tokenizer trained on language-specific HPLT data.
[The training code](https://github.com/hplt-project/HPLT-WP4).
[The training statistics of all runs](https://api.wandb.ai/links/ltg/kduj7mjn)
## Example usage (tested with `transformers==4.46.1` and `tokenizers==0.20.1`)
This model currently needs a custom wrapper from `modeling_ltgbert.py`, you should therefore load the model with `trust_remote_code=True`.
```python
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("HPLT/hplt_bert_base_2_0_ita-Latn")
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_ita-Latn", trust_remote_code=True)
mask_id = tokenizer.convert_tokens_to_ids("[MASK]")
input_text = tokenizer("It's a beautiful[MASK].", return_tensors="pt")
output_p = model(**input_text)
output_text = torch.where(input_text.input_ids == mask_id, output_p.logits.argmax(-1), input_text.input_ids)
# should output: '[CLS] It's a beautiful place.[SEP]'
print(tokenizer.decode(output_text[0].tolist(), clean_up_tokenization_spaces=True))
```
The following classes are currently implemented: `AutoModel`, `AutoModelMaskedLM`, `AutoModelForSequenceClassification`, `AutoModelForTokenClassification`, `AutoModelForQuestionAnswering` and `AutoModeltForMultipleChoice`.
## Intermediate checkpoints
We are releasing 10 intermediate checkpoints for each model at intervals of every 3125 training steps in separate branches. The naming convention is `stepXXX`: for example, `step18750`.
You can load a specific model revision with `transformers` using the argument `revision`:
```python
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_ita-Latn", revision="step21875", trust_remote_code=True)
```
You can access all the revisions for the models with the following code:
```python
from huggingface_hub import list_repo_refs
out = list_repo_refs("HPLT/hplt_bert_base_2_0_ita-Latn")
print([b.name for b in out.branches])
```
## Cite us
```bibtex
@inproceedings{samuel-etal-2023-trained,
title = "Trained on 100 million words and still in shape: {BERT} meets {B}ritish {N}ational {C}orpus",
author = "Samuel, David and
Kutuzov, Andrey and
{\O}vrelid, Lilja and
Velldal, Erik",
editor = "Vlachos, Andreas and
Augenstein, Isabelle",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2023",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-eacl.146",
doi = "10.18653/v1/2023.findings-eacl.146",
pages = "1954--1974"
}
```
```bibtex
@misc{burchell2025expandedmassivemultilingualdataset,
title={An Expanded Massive Multilingual Dataset for High-Performance Language Technologies},
author={Laurie Burchell and Ona de Gibert and Nikolay Arefyev and Mikko Aulamo and Marta Bañón and Pinzhen Chen and Mariia Fedorova and Liane Guillou and Barry Haddow and Jan Hajič and Jindřich Helcl and Erik Henriksson and Mateusz Klimaszewski and Ville Komulainen and Andrey Kutuzov and Joona Kytöniemi and Veronika Laippala and Petter Mæhlum and Bhavitvya Malik and Farrokh Mehryary and Vladislav Mikhailov and Nikita Moghe and Amanda Myntti and Dayyán O'Brien and Stephan Oepen and Proyag Pal and Jousia Piha and Sampo Pyysalo and Gema Ramírez-Sánchez and David Samuel and Pavel Stepachev and Jörg Tiedemann and Dušan Variš and Tereza Vojtěchová and Jaume Zaragoza-Bernabeu},
year={2025},
eprint={2503.10267},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2503.10267},
}
```
|
HPLT/hplt_bert_base_2_0_fao-Latn | HPLT | 2025-05-29T13:20:50Z | 0 | 0 | null | [
"pytorch",
"BERT",
"HPLT",
"encoder",
"fill-mask",
"custom_code",
"fo",
"fao",
"dataset:HPLT/HPLT2.0_cleaned",
"arxiv:2503.10267",
"license:apache-2.0",
"region:us"
] | fill-mask | 2025-02-22T22:42:47Z | ---
language:
- fo
- fao
inference: false
tags:
- BERT
- HPLT
- encoder
license: apache-2.0
datasets:
- HPLT/HPLT2.0_cleaned
pipeline_tag: fill-mask
---
# HPLT v2.0 BERT for Faroese
<img src="https://hplt-project.org/_next/static/media/logo-hplt.d5e16ca5.svg" width=12.5%>
This is one of the encoder-only monolingual language models trained as a second release by the [HPLT project](https://hplt-project.org/).
It is a so called masked language model. In particular, we used the modification of the classic BERT model named [LTG-BERT](https://aclanthology.org/2023.findings-eacl.146/).
We present monolingual LTG-BERT models for more than 50 languages out of 191 total in the [HPLT v2.0 dataset](https://hplt-project.org/datasets/v2.0).
All the HPLT encoder-only models use the same hyper-parameters, roughly following the BERT-base setup:
- hidden size: 768
- attention heads: 12
- layers: 12
- vocabulary size: 32768
Every model uses its own tokenizer trained on language-specific HPLT data.
[The training code](https://github.com/hplt-project/HPLT-WP4).
[The training statistics of all runs](https://api.wandb.ai/links/ltg/kduj7mjn)
## Example usage (tested with `transformers==4.46.1` and `tokenizers==0.20.1`)
This model currently needs a custom wrapper from `modeling_ltgbert.py`, you should therefore load the model with `trust_remote_code=True`.
```python
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("HPLT/hplt_bert_base_2_0_fao-Latn")
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_fao-Latn", trust_remote_code=True)
mask_id = tokenizer.convert_tokens_to_ids("[MASK]")
input_text = tokenizer("It's a beautiful[MASK].", return_tensors="pt")
output_p = model(**input_text)
output_text = torch.where(input_text.input_ids == mask_id, output_p.logits.argmax(-1), input_text.input_ids)
# should output: '[CLS] It's a beautiful place.[SEP]'
print(tokenizer.decode(output_text[0].tolist(), clean_up_tokenization_spaces=True))
```
The following classes are currently implemented: `AutoModel`, `AutoModelMaskedLM`, `AutoModelForSequenceClassification`, `AutoModelForTokenClassification`, `AutoModelForQuestionAnswering` and `AutoModeltForMultipleChoice`.
## Intermediate checkpoints
We are releasing 10 intermediate checkpoints for each model at intervals of every 3125 training steps in separate branches. The naming convention is `stepXXX`: for example, `step18750`.
You can load a specific model revision with `transformers` using the argument `revision`:
```python
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_fao-Latn", revision="step21875", trust_remote_code=True)
```
You can access all the revisions for the models with the following code:
```python
from huggingface_hub import list_repo_refs
out = list_repo_refs("HPLT/hplt_bert_base_2_0_fao-Latn")
print([b.name for b in out.branches])
```
## Cite us
```bibtex
@inproceedings{samuel-etal-2023-trained,
title = "Trained on 100 million words and still in shape: {BERT} meets {B}ritish {N}ational {C}orpus",
author = "Samuel, David and
Kutuzov, Andrey and
{\O}vrelid, Lilja and
Velldal, Erik",
editor = "Vlachos, Andreas and
Augenstein, Isabelle",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2023",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-eacl.146",
doi = "10.18653/v1/2023.findings-eacl.146",
pages = "1954--1974"
}
```
```bibtex
@misc{burchell2025expandedmassivemultilingualdataset,
title={An Expanded Massive Multilingual Dataset for High-Performance Language Technologies},
author={Laurie Burchell and Ona de Gibert and Nikolay Arefyev and Mikko Aulamo and Marta Bañón and Pinzhen Chen and Mariia Fedorova and Liane Guillou and Barry Haddow and Jan Hajič and Jindřich Helcl and Erik Henriksson and Mateusz Klimaszewski and Ville Komulainen and Andrey Kutuzov and Joona Kytöniemi and Veronika Laippala and Petter Mæhlum and Bhavitvya Malik and Farrokh Mehryary and Vladislav Mikhailov and Nikita Moghe and Amanda Myntti and Dayyán O'Brien and Stephan Oepen and Proyag Pal and Jousia Piha and Sampo Pyysalo and Gema Ramírez-Sánchez and David Samuel and Pavel Stepachev and Jörg Tiedemann and Dušan Variš and Tereza Vojtěchová and Jaume Zaragoza-Bernabeu},
year={2025},
eprint={2503.10267},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2503.10267},
}
```
|
HPLT/hplt_bert_base_2_0_eng-Latn | HPLT | 2025-05-29T13:20:00Z | 103 | 1 | null | [
"pytorch",
"BERT",
"HPLT",
"encoder",
"fill-mask",
"custom_code",
"en",
"eng",
"dataset:HPLT/HPLT2.0_cleaned",
"arxiv:2503.10267",
"license:apache-2.0",
"region:us"
] | fill-mask | 2025-02-22T02:02:10Z | ---
language:
- en
- eng
inference: false
tags:
- BERT
- HPLT
- encoder
license: apache-2.0
datasets:
- HPLT/HPLT2.0_cleaned
pipeline_tag: fill-mask
---
# HPLT v2.0 BERT for English
<img src="https://hplt-project.org/_next/static/media/logo-hplt.d5e16ca5.svg" width=12.5%>
This is one of the encoder-only monolingual language models trained as a second release by the [HPLT project](https://hplt-project.org/).
It is a so called masked language model. In particular, we used the modification of the classic BERT model named [LTG-BERT](https://aclanthology.org/2023.findings-eacl.146/).
We present monolingual LTG-BERT models for more than 50 languages out of 191 total in the [HPLT v2.0 dataset](https://hplt-project.org/datasets/v2.0) (*52* models total).
All the HPLT encoder-only models use the same hyper-parameters, roughly following the BERT-base setup:
- hidden size: 768
- attention heads: 12
- layers: 12
- vocabulary size: 32768
Every model uses its own tokenizer trained on language-specific HPLT data.
[The training code](https://github.com/hplt-project/HPLT-WP4).
[The training statistics of all runs](https://api.wandb.ai/links/ltg/kduj7mjn)
## Example usage (tested with `transformers==4.46.1` and `tokenizers==0.20.1`)
This model currently needs a custom wrapper from `modeling_ltgbert.py`, you should therefore load the model with `trust_remote_code=True`.
```python
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("HPLT/hplt_bert_base_2_0_eng-Latn")
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_eng-Latn", trust_remote_code=True)
mask_id = tokenizer.convert_tokens_to_ids("[MASK]")
input_text = tokenizer("It's a beautiful[MASK].", return_tensors="pt")
output_p = model(**input_text)
output_text = torch.where(input_text.input_ids == mask_id, output_p.logits.argmax(-1), input_text.input_ids)
# should output: '[CLS] It's a beautiful place.[SEP]'
print(tokenizer.decode(output_text[0].tolist(), clean_up_tokenization_spaces=True))
```
The following classes are currently implemented: `AutoModel`, `AutoModelMaskedLM`, `AutoModelForSequenceClassification`, `AutoModelForTokenClassification`, `AutoModelForQuestionAnswering` and `AutoModeltForMultipleChoice`.
## Intermediate checkpoints
We are releasing 10 intermediate checkpoints for each model at intervals of every 3125 training steps in separate branches. The naming convention is `stepXXX`: for example, `step18750`.
You can load a specific model revision with `transformers` using the argument `revision`:
```python
model = AutoModelForMaskedLM.from_pretrained("HPLT/hplt_bert_base_2_0_eng-Latn", revision="step21875", trust_remote_code=True)
```
You can access all the revisions for the models with the following code:
```python
from huggingface_hub import list_repo_refs
out = list_repo_refs("HPLT/hplt_bert_base_2_0_eng-Latn")
print([b.name for b in out.branches])
```
## Cite us
```bibtex
@inproceedings{samuel-etal-2023-trained,
title = "Trained on 100 million words and still in shape: {BERT} meets {B}ritish {N}ational {C}orpus",
author = "Samuel, David and
Kutuzov, Andrey and
{\O}vrelid, Lilja and
Velldal, Erik",
editor = "Vlachos, Andreas and
Augenstein, Isabelle",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2023",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-eacl.146",
doi = "10.18653/v1/2023.findings-eacl.146",
pages = "1954--1974"
})
```
```bibtex
@misc{burchell2025expandedmassivemultilingualdataset,
title={An Expanded Massive Multilingual Dataset for High-Performance Language Technologies},
author={Laurie Burchell and Ona de Gibert and Nikolay Arefyev and Mikko Aulamo and Marta Bañón and Pinzhen Chen and Mariia Fedorova and Liane Guillou and Barry Haddow and Jan Hajič and Jindřich Helcl and Erik Henriksson and Mateusz Klimaszewski and Ville Komulainen and Andrey Kutuzov and Joona Kytöniemi and Veronika Laippala and Petter Mæhlum and Bhavitvya Malik and Farrokh Mehryary and Vladislav Mikhailov and Nikita Moghe and Amanda Myntti and Dayyán O'Brien and Stephan Oepen and Proyag Pal and Jousia Piha and Sampo Pyysalo and Gema Ramírez-Sánchez and David Samuel and Pavel Stepachev and Jörg Tiedemann and Dušan Variš and Tereza Vojtěchová and Jaume Zaragoza-Bernabeu},
year={2025},
eprint={2503.10267},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2503.10267},
}
```
|
gulshany01/qwen2_0.5b_instruct_invoice_parser_v2 | gulshany01 | 2025-05-29T13:13:20Z | 0 | 0 | transformers | [
"transformers",
"gguf",
"qwen2",
"text-generation-inference",
"unsloth",
"en",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-05-29T13:12:49Z | ---
base_model: unsloth/qwen2-0.5b-instruct-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- qwen2
- gguf
license: apache-2.0
language:
- en
---
# Uploaded model
- **Developed by:** gulshany01
- **License:** apache-2.0
- **Finetuned from model :** unsloth/qwen2-0.5b-instruct-bnb-4bit
This qwen2 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
Electricboogaloo6/ppo-LunarLander-v2 | Electricboogaloo6 | 2025-05-29T13:09:45Z | 0 | 0 | stable-baselines3 | [
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] | reinforcement-learning | 2025-05-29T13:09:26Z | ---
library_name: stable-baselines3
tags:
- LunarLander-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: 257.30 +/- 43.64
name: mean_reward
verified: false
---
# **PPO** Agent playing **LunarLander-v2**
This is a trained model of a **PPO** agent playing **LunarLander-v2**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
Anish20011/Anish | Anish20011 | 2025-05-29T12:59:39Z | 0 | 0 | null | [
"license:apache-2.0",
"region:us"
] | null | 2025-05-29T12:59:39Z | ---
license: apache-2.0
---
|
Hsianchengfun/meta-1B-KD_lora_single_device-epoch_0 | Hsianchengfun | 2025-05-29T12:55:07Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"llama",
"text-generation",
"facebook",
"meta",
"pytorch",
"llama-3",
"conversational",
"en",
"de",
"fr",
"it",
"pt",
"hi",
"es",
"th",
"arxiv:2204.05149",
"arxiv:2405.16406",
"license:llama3.2",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-29T12:53:42Z | ---
language:
- en
- de
- fr
- it
- pt
- hi
- es
- th
library_name: transformers
pipeline_tag: text-generation
tags:
- facebook
- meta
- pytorch
- llama
- llama-3
license: llama3.2
extra_gated_prompt: >-
### LLAMA 3.2 COMMUNITY LICENSE AGREEMENT
Llama 3.2 Version Release Date: September 25, 2024
“Agreement” means the terms and conditions for use, reproduction, distribution
and modification of the Llama Materials set forth herein.
“Documentation” means the specifications, manuals and documentation accompanying Llama 3.2
distributed by Meta at https://llama.meta.com/doc/overview.
“Licensee” or “you” means you, or your employer or any other person or entity (if you are
entering into this Agreement on such person or entity’s behalf), of the age required under
applicable laws, rules or regulations to provide legal consent and that has legal authority
to bind your employer or such other person or entity if you are entering in this Agreement
on their behalf.
“Llama 3.2” means the foundational large language models and software and algorithms, including
machine-learning model code, trained model weights, inference-enabling code, training-enabling code,
fine-tuning enabling code and other elements of the foregoing distributed by Meta at
https://www.llama.com/llama-downloads.
“Llama Materials” means, collectively, Meta’s proprietary Llama 3.2 and Documentation (and
any portion thereof) made available under this Agreement.
“Meta” or “we” means Meta Platforms Ireland Limited (if you are located in or,
if you are an entity, your principal place of business is in the EEA or Switzerland)
and Meta Platforms, Inc. (if you are located outside of the EEA or Switzerland).
By clicking “I Accept” below or by using or distributing any portion or element of the Llama Materials,
you agree to be bound by this Agreement.
1. License Rights and Redistribution.
a. Grant of Rights. You are granted a non-exclusive, worldwide,
non-transferable and royalty-free limited license under Meta’s intellectual property or other rights
owned by Meta embodied in the Llama Materials to use, reproduce, distribute, copy, create derivative works
of, and make modifications to the Llama Materials.
b. Redistribution and Use.
i. If you distribute or make available the Llama Materials (or any derivative works thereof),
or a product or service (including another AI model) that contains any of them, you shall (A) provide
a copy of this Agreement with any such Llama Materials; and (B) prominently display “Built with Llama”
on a related website, user interface, blogpost, about page, or product documentation. If you use the
Llama Materials or any outputs or results of the Llama Materials to create, train, fine tune, or
otherwise improve an AI model, which is distributed or made available, you shall also include “Llama”
at the beginning of any such AI model name.
ii. If you receive Llama Materials, or any derivative works thereof, from a Licensee as part
of an integrated end user product, then Section 2 of this Agreement will not apply to you.
iii. You must retain in all copies of the Llama Materials that you distribute the
following attribution notice within a “Notice” text file distributed as a part of such copies:
“Llama 3.2 is licensed under the Llama 3.2 Community License, Copyright © Meta Platforms,
Inc. All Rights Reserved.”
iv. Your use of the Llama Materials must comply with applicable laws and regulations
(including trade compliance laws and regulations) and adhere to the Acceptable Use Policy for
the Llama Materials (available at https://www.llama.com/llama3_2/use-policy), which is hereby
incorporated by reference into this Agreement.
2. Additional Commercial Terms. If, on the Llama 3.2 version release date, the monthly active users
of the products or services made available by or for Licensee, or Licensee’s affiliates,
is greater than 700 million monthly active users in the preceding calendar month, you must request
a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to
exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights.
3. Disclaimer of Warranty. UNLESS REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS AND ANY OUTPUT AND
RESULTS THEREFROM ARE PROVIDED ON AN “AS IS” BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS
ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES
OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE SOLELY RESPONSIBLE
FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED
WITH YOUR USE OF THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS.
4. Limitation of Liability. IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY,
WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING OUT OF THIS AGREEMENT,
FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN
IF META OR ITS AFFILIATES HAVE BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.
5. Intellectual Property.
a. No trademark licenses are granted under this Agreement, and in connection with the Llama Materials,
neither Meta nor Licensee may use any name or mark owned by or associated with the other or any of its affiliates,
except as required for reasonable and customary use in describing and redistributing the Llama Materials or as
set forth in this Section 5(a). Meta hereby grants you a license to use “Llama” (the “Mark”) solely as required
to comply with the last sentence of Section 1.b.i. You will comply with Meta’s brand guidelines (currently accessible
at https://about.meta.com/brand/resources/meta/company-brand/). All goodwill arising out of your use of the Mark
will inure to the benefit of Meta.
b. Subject to Meta’s ownership of Llama Materials and derivatives made by or for Meta, with respect to any
derivative works and modifications of the Llama Materials that are made by you, as between you and Meta,
you are and will be the owner of such derivative works and modifications.
c. If you institute litigation or other proceedings against Meta or any entity (including a cross-claim or
counterclaim in a lawsuit) alleging that the Llama Materials or Llama 3.2 outputs or results, or any portion
of any of the foregoing, constitutes infringement of intellectual property or other rights owned or licensable
by you, then any licenses granted to you under this Agreement shall terminate as of the date such litigation or
claim is filed or instituted. You will indemnify and hold harmless Meta from and against any claim by any third
party arising out of or related to your use or distribution of the Llama Materials.
6. Term and Termination. The term of this Agreement will commence upon your acceptance of this Agreement or access
to the Llama Materials and will continue in full force and effect until terminated in accordance with the terms
and conditions herein. Meta may terminate this Agreement if you are in breach of any term or condition of this
Agreement. Upon termination of this Agreement, you shall delete and cease use of the Llama Materials. Sections 3,
4 and 7 shall survive the termination of this Agreement.
7. Governing Law and Jurisdiction. This Agreement will be governed and construed under the laws of the State of
California without regard to choice of law principles, and the UN Convention on Contracts for the International
Sale of Goods does not apply to this Agreement. The courts of California shall have exclusive jurisdiction of
any dispute arising out of this Agreement.
### Llama 3.2 Acceptable Use Policy
Meta is committed to promoting safe and fair use of its tools and features, including Llama 3.2.
If you access or use Llama 3.2, you agree to this Acceptable Use Policy (“**Policy**”).
The most recent copy of this policy can be found at
[https://www.llama.com/llama3_2/use-policy](https://www.llama.com/llama3_2/use-policy).
#### Prohibited Uses
We want everyone to use Llama 3.2 safely and responsibly. You agree you will not use, or allow others to use, Llama 3.2 to:
1. Violate the law or others’ rights, including to:
1. Engage in, promote, generate, contribute to, encourage, plan, incite, or further illegal or unlawful activity or content, such as:
1. Violence or terrorism
2. Exploitation or harm to children, including the solicitation, creation, acquisition, or dissemination of child exploitative content or failure to report Child Sexual Abuse Material
3. Human trafficking, exploitation, and sexual violence
4. The illegal distribution of information or materials to minors, including obscene materials, or failure to employ legally required age-gating in connection with such information or materials.
5. Sexual solicitation
6. Any other criminal activity
1. Engage in, promote, incite, or facilitate the harassment, abuse, threatening, or bullying of individuals or groups of individuals
2. Engage in, promote, incite, or facilitate discrimination or other unlawful or harmful conduct in the provision of employment, employment benefits, credit, housing, other economic benefits, or other essential goods and services
3. Engage in the unauthorized or unlicensed practice of any profession including, but not limited to, financial, legal, medical/health, or related professional practices
4. Collect, process, disclose, generate, or infer private or sensitive information about individuals, including information about individuals’ identity, health, or demographic information, unless you have obtained the right to do so in accordance with applicable law
5. Engage in or facilitate any action or generate any content that infringes, misappropriates, or otherwise violates any third-party rights, including the outputs or results of any products or services using the Llama Materials
6. Create, generate, or facilitate the creation of malicious code, malware, computer viruses or do anything else that could disable, overburden, interfere with or impair the proper working, integrity, operation or appearance of a website or computer system
7. Engage in any action, or facilitate any action, to intentionally circumvent or remove usage restrictions or other safety measures, or to enable functionality disabled by Meta
2. Engage in, promote, incite, facilitate, or assist in the planning or development of activities that present a risk of death or bodily harm to individuals, including use of Llama 3.2 related to the following:
8. Military, warfare, nuclear industries or applications, espionage, use for materials or activities that are subject to the International Traffic Arms Regulations (ITAR) maintained by the United States Department of State or to the U.S. Biological Weapons Anti-Terrorism Act of 1989 or the Chemical Weapons Convention Implementation Act of 1997
9. Guns and illegal weapons (including weapon development)
10. Illegal drugs and regulated/controlled substances
11. Operation of critical infrastructure, transportation technologies, or heavy machinery
12. Self-harm or harm to others, including suicide, cutting, and eating disorders
13. Any content intended to incite or promote violence, abuse, or any infliction of bodily harm to an individual
3. Intentionally deceive or mislead others, including use of Llama 3.2 related to the following:
14. Generating, promoting, or furthering fraud or the creation or promotion of disinformation
15. Generating, promoting, or furthering defamatory content, including the creation of defamatory statements, images, or other content
16. Generating, promoting, or further distributing spam
17. Impersonating another individual without consent, authorization, or legal right
18. Representing that the use of Llama 3.2 or outputs are human-generated
19. Generating or facilitating false online engagement, including fake reviews and other means of fake online engagement
4. Fail to appropriately disclose to end users any known dangers of your AI system
5. Interact with third party tools, models, or software designed to generate unlawful content or engage in unlawful or harmful conduct and/or represent that the outputs of such tools, models, or software are associated with Meta or Llama 3.2
With respect to any multimodal models included in Llama 3.2, the rights granted under Section 1(a) of the Llama 3.2 Community License Agreement are not being granted to you if you are an individual domiciled in, or a company with a principal place of business in, the European Union. This restriction does not apply to end users of a product or service that incorporates any such multimodal models.
Please report any violation of this Policy, software “bug,” or other problems that could lead to a violation of this Policy through one of the following means:
* Reporting issues with the model: [https://github.com/meta-llama/llama-models/issues](https://l.workplace.com/l.php?u=https%3A%2F%2Fgithub.com%2Fmeta-llama%2Fllama-models%2Fissues&h=AT0qV8W9BFT6NwihiOHRuKYQM_UnkzN_NmHMy91OT55gkLpgi4kQupHUl0ssR4dQsIQ8n3tfd0vtkobvsEvt1l4Ic6GXI2EeuHV8N08OG2WnbAmm0FL4ObkazC6G_256vN0lN9DsykCvCqGZ)
* Reporting risky content generated by the model: [developers.facebook.com/llama_output_feedback](http://developers.facebook.com/llama_output_feedback)
* Reporting bugs and security concerns: [facebook.com/whitehat/info](http://facebook.com/whitehat/info)
* Reporting violations of the Acceptable Use Policy or unlicensed uses of Llama 3.2: [email protected]
extra_gated_fields:
First Name: text
Last Name: text
Date of birth: date_picker
Country: country
Affiliation: text
Job title:
type: select
options:
- Student
- Research Graduate
- AI researcher
- AI developer/engineer
- Reporter
- Other
geo: ip_location
By clicking Submit below I accept the terms of the license and acknowledge that the information I provide will be collected stored processed and shared in accordance with the Meta Privacy Policy: checkbox
extra_gated_description: >-
The information you provide will be collected, stored, processed and shared in
accordance with the [Meta Privacy
Policy](https://www.facebook.com/privacy/policy/).
extra_gated_button_content: Submit
---
## Model Information
The Llama 3.2 collection of multilingual large language models (LLMs) is a collection of pretrained and instruction-tuned generative models in 1B and 3B sizes (text in/text out). The Llama 3.2 instruction-tuned text only models are optimized for multilingual dialogue use cases, including agentic retrieval and summarization tasks. They outperform many of the available open source and closed chat models on common industry benchmarks.
**Model Developer:** Meta
**Model Architecture:** Llama 3.2 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.
| | Training Data | Params | Input modalities | Output modalities | Context Length | GQA | Shared Embeddings | Token count | Knowledge cutoff |
| :---- | :---- | :---- | :---- | :---- | :---- | :---- | :---- | :---- | :---- |
| Llama 3.2 (text only) | A new mix of publicly available online data. | 1B (1.23B) | Multilingual Text | Multilingual Text and code | 128k | Yes | Yes | Up to 9T tokens | December 2023 |
| | | 3B (3.21B) | Multilingual Text | Multilingual Text and code | | | | | |
| Llama 3.2 Quantized (text only) | A new mix of publicly available online data. | 1B (1.23B) | Multilingual Text | Multilingual Text and code | 8k | Yes | Yes | Up to 9T tokens | December 2023 |
| | | 3B (3.21B) | Multilingual Text | Multilingual Text and code | | | | | |
**Supported Languages:** English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai are officially supported. Llama 3.2 has been trained on a broader collection of languages than these 8 supported languages. Developers may fine-tune Llama 3.2 models for languages beyond these supported languages, provided they comply with the Llama 3.2 Community License and the Acceptable Use Policy. Developers are always expected to ensure that their deployments, including those that involve additional languages, are completed safely and responsibly.
**Llama 3.2 Model Family:** Token counts refer to pretraining data only. All model versions use Grouped-Query Attention (GQA) for improved inference scalability.
**Model Release Date:** Sept 25, 2024
**Status:** This is a static model trained on an offline dataset. Future versions may be released that improve model capabilities and safety.
**License:** Use of Llama 3.2 is governed by the [Llama 3.2 Community License](https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/LICENSE) (a custom, commercial license agreement).
**Feedback:** Instructions on how to provide feedback or comments on the model can be found in the Llama Models [README](https://github.com/meta-llama/llama-models/blob/main/README.md). For more technical information about generation parameters and recipes for how to use Llama 3.2 in applications, please go [here](https://github.com/meta-llama/llama-recipes).
## Intended Use
**Intended Use Cases:** Llama 3.2 is intended for commercial and research use in multiple languages. Instruction tuned text only models are intended for assistant-like chat and agentic applications like knowledge retrieval and summarization, mobile AI powered writing assistants and query and prompt rewriting. Pretrained models can be adapted for a variety of additional natural language generation tasks. Similarly, quantized models can be adapted for a variety of on-device use-cases with limited compute resources.
**Out of Scope:** Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3.2 Community License. Use in languages beyond those explicitly referenced as supported in this model card.
## How to use
This repository contains two versions of Llama-3.2-1B-Instruct, for use with transformers and with the original `llama` codebase.
### Use with transformers
Starting with `transformers >= 4.43.0` onward, you can run conversational inference using the Transformers `pipeline` abstraction or by leveraging the Auto classes with the `generate()` function.
Make sure to update your transformers installation via `pip install --upgrade transformers`.
```python
import torch
from transformers import pipeline
model_id = "meta-llama/Llama-3.2-1B-Instruct"
pipe = pipeline(
"text-generation",
model=model_id,
torch_dtype=torch.bfloat16,
device_map="auto",
)
messages = [
{"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
{"role": "user", "content": "Who are you?"},
]
outputs = pipe(
messages,
max_new_tokens=256,
)
print(outputs[0]["generated_text"][-1])
```
Note: You can also find detailed recipes on how to use the model locally, with `torch.compile()`, assisted generations, quantised and more at [`huggingface-llama-recipes`](https://github.com/huggingface/huggingface-llama-recipes)
### Use with `llama`
Please, follow the instructions in the [repository](https://github.com/meta-llama/llama)
To download Original checkpoints, see the example command below leveraging `huggingface-cli`:
```
huggingface-cli download meta-llama/Llama-3.2-1B-Instruct --include "original/*" --local-dir Llama-3.2-1B-Instruct
```
## Hardware and Software
**Training Factors:** We used custom training libraries, Meta's custom built GPU cluster, and production infrastructure for pretraining. Fine-tuning, quantization, annotation, and evaluation were also performed on production infrastructure.
**Training Energy Use:** Training utilized a cumulative of **916k** GPU hours of computation on H100-80GB (TDP of 700W) type hardware, per the table below. Training time is the total GPU time required for training each model and power consumption is the peak power capacity per GPU device used, adjusted for power usage efficiency.
**Training Greenhouse Gas Emissions:** Estimated total location-based greenhouse gas emissions were **240** tons CO2eq for training. Since 2020, Meta has maintained net zero greenhouse gas emissions in its global operations and matched 100% of its electricity use with renewable energy; therefore, the total market-based greenhouse gas emissions for training were 0 tons CO2eq.
| | Training Time (GPU hours) | Logit Generation Time (GPU Hours) | Training Power Consumption (W) | Training Location-Based Greenhouse Gas Emissions (tons CO2eq) | Training Market-Based Greenhouse Gas Emissions (tons CO2eq) |
| :---- | :---: | ----- | :---: | :---: | :---: |
| Llama 3.2 1B | 370k | \- | 700 | 107 | 0 |
| Llama 3.2 3B | 460k | \- | 700 | 133 | 0 |
| Llama 3.2 1B SpinQuant | 1.7 | 0 | 700 | *Negligible*\*\* | 0 |
| Llama 3.2 3B SpinQuant | 2.4 | 0 | 700 | *Negligible*\*\* | 0 |
| Llama 3.2 1B QLora | 1.3k | 0 | 700 | 0.381 | 0 |
| Llama 3.2 3B QLora | 1.6k | 0 | 700 | 0.461 | 0 |
| Total | 833k | 86k | | 240 | 0 |
\*\* The location-based CO2e emissions of Llama 3.2 1B SpinQuant and Llama 3.2 3B SpinQuant are less than 0.001 metric tonnes each. This is due to the minimal training GPU hours that are required.
The methodology used to determine training energy use and greenhouse gas emissions can be found [here](https://arxiv.org/pdf/2204.05149). Since Meta is openly releasing these models, the training energy use and greenhouse gas emissions will not be incurred by others.
## Training Data
**Overview:** Llama 3.2 was pretrained on up to 9 trillion tokens of data from publicly available sources. For the 1B and 3B Llama 3.2 models, we incorporated logits from the Llama 3.1 8B and 70B models into the pretraining stage of the model development, where outputs (logits) from these larger models were used as token-level targets. Knowledge distillation was used after pruning to recover performance. In post-training we used a similar recipe as Llama 3.1 and produced final chat models by doing several rounds of alignment on top of the pre-trained model. Each round involved Supervised Fine-Tuning (SFT), Rejection Sampling (RS), and Direct Preference Optimization (DPO).
**Data Freshness:** The pretraining data has a cutoff of December 2023\.
## Quantization
### Quantization Scheme
We designed the current quantization scheme with the [PyTorch’s ExecuTorch](https://github.com/pytorch/executorch) inference framework and Arm CPU backend in mind, taking into account metrics including model quality, prefill/decoding speed, and memory footprint. Our quantization scheme involves three parts:
- All linear layers in all transformer blocks are quantized to a 4-bit groupwise scheme (with a group size of 32) for weights and 8-bit per-token dynamic quantization for activations.
- The classification layer is quantized to 8-bit per-channel for weight and 8-bit per token dynamic quantization for activation.
- Similar to classification layer, an 8-bit per channel quantization is used for embedding layer.
### Quantization-Aware Training and LoRA
The quantization-aware training (QAT) with low-rank adaptation (LoRA) models went through only post-training stages, using the same data as the full precision models. To initialize QAT, we utilize BF16 Llama 3.2 model checkpoints obtained after supervised fine-tuning (SFT) and perform an additional full round of SFT training with QAT. We then freeze the backbone of the QAT model and perform another round of SFT with LoRA adaptors applied to all layers within the transformer block. Meanwhile, the LoRA adaptors' weights and activations are maintained in BF16. Because our approach is similar to QLoRA of Dettmers et al., (2023) (i.e., quantization followed by LoRA adapters), we refer this method as QLoRA. Finally, we fine-tune the resulting model (both backbone and LoRA adaptors) using direct preference optimization (DPO).
### SpinQuant
[SpinQuant](https://arxiv.org/abs/2405.16406) was applied, together with generative post-training quantization (GPTQ). For the SpinQuant rotation matrix fine-tuning, we optimized for 100 iterations, using 800 samples with sequence-length 2048 from the WikiText 2 dataset. For GPTQ, we used 128 samples from the same dataset with the same sequence-length.
## Benchmarks \- English Text
In this section, we report the results for Llama 3.2 models on standard automatic benchmarks. For all these evaluations, we used our internal evaluations library.
### Base Pretrained Models
| Category | Benchmark | \# Shots | Metric | Llama 3.2 1B | Llama 3.2 3B | Llama 3.1 8B |
| ----- | ----- | :---: | :---: | :---: | :---: | :---: |
| General | MMLU | 5 | macro\_avg/acc\_char | 32.2 | 58 | 66.7 |
| | AGIEval English | 3-5 | average/acc\_char | 23.3 | 39.2 | 47.8 |
| | ARC-Challenge | 25 | acc\_char | 32.8 | 69.1 | 79.7 |
| Reading comprehension | SQuAD | 1 | em | 49.2 | 67.7 | 77 |
| | QuAC (F1) | 1 | f1 | 37.9 | 42.9 | 44.9 |
| | DROP (F1) | 3 | f1 | 28.0 | 45.2 | 59.5 |
| Long Context | Needle in Haystack | 0 | em | 96.8 | 1 | 1 |
### Instruction Tuned Models
| Capability | | Benchmark | \# Shots | Metric | Llama 3.2 1B bf16 | Llama 3.2 1B Vanilla PTQ\*\* | Llama 3.2 1B Spin Quant | Llama 3.2 1B QLoRA | Llama 3.2 3B bf16 | Llama 3.2 3B Vanilla PTQ\*\* | Llama 3.2 3B Spin Quant | Llama 3.2 3B QLoRA | Llama 3.1 8B |
| :---: | ----- | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
| General | | MMLU | 5 | macro\_avg/acc | 49.3 | 43.3 | 47.3 | 49.0 | 63.4 | 60.5 | 62 | 62.4 | 69.4 |
| Re-writing | | Open-rewrite eval | 0 | micro\_avg/rougeL | 41.6 | 39.2 | 40.9 | 41.2 | 40.1 | 40.3 | 40.8 | 40.7 | 40.9 |
| Summarization | | TLDR9+ (test) | 1 | rougeL | 16.8 | 14.9 | 16.7 | 16.8 | 19.0 | 19.1 | 19.2 | 19.1 | 17.2 |
| Instruction following | | IFEval | 0 | Avg(Prompt/Instruction acc Loose/Strict) | 59.5 | 51.5 | 58.4 | 55.6 | 77.4 | 73.9 | 73.5 | 75.9 | 80.4 |
| Math | | GSM8K (CoT) | 8 | em\_maj1@1 | 44.4 | 33.1 | 40.6 | 46.5 | 77.7 | 72.9 | 75.7 | 77.9 | 84.5 |
| | | MATH (CoT) | 0 | final\_em | 30.6 | 20.5 | 25.3 | 31.0 | 48.0 | 44.2 | 45.3 | 49.2 | 51.9 |
| Reasoning | | ARC-C | 0 | acc | 59.4 | 54.3 | 57 | 60.7 | 78.6 | 75.6 | 77.6 | 77.6 | 83.4 |
| | | GPQA | 0 | acc | 27.2 | 25.9 | 26.3 | 25.9 | 32.8 | 32.8 | 31.7 | 33.9 | 32.8 |
| | | Hellaswag | 0 | acc | 41.2 | 38.1 | 41.3 | 41.5 | 69.8 | 66.3 | 68 | 66.3 | 78.7 |
| Tool Use | | BFCL V2 | 0 | acc | 25.7 | 14.3 | 15.9 | 23.7 | 67.0 | 53.4 | 60.1 | 63.5 | 67.1 |
| | | Nexus | 0 | macro\_avg/acc | 13.5 | 5.2 | 9.6 | 12.5 | 34.3 | 32.4 | 31.5 | 30.1 | 38.5 |
| Long Context | | InfiniteBench/En.QA | 0 | longbook\_qa/f1 | 20.3 | N/A | N/A | N/A | 19.8 | N/A | N/A | N/A | 27.3 |
| | | InfiniteBench/En.MC | 0 | longbook\_choice/acc | 38.0 | N/A | N/A | N/A | 63.3 | N/A | N/A | N/A | 72.2 |
| | | NIH/Multi-needle | 0 | recall | 75.0 | N/A | N/A | N/A | 84.7 | N/A | N/A | N/A | 98.8 |
| Multilingual | | MGSM (CoT) | 0 | em | 24.5 | 13.7 | 18.2 | 24.4 | 58.2 | 48.9 | 54.3 | 56.8 | 68.9 |
\*\*for comparison purposes only. Model not released.
### Multilingual Benchmarks
| Category | Benchmark | Language | Llama 3.2 1B | Llama 3.2 1B Vanilla PTQ\*\* | Llama 3.2 1B Spin Quant | Llama 3.2 1B QLoRA | Llama 3.2 3B | Llama 3.2 3B Vanilla PTQ\*\* | Llama 3.2 3B Spin Quant | Llama 3.2 3B QLoRA | Llama 3.1 8B |
| :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
| General | MMLU (5-shot, macro_avg/acc) | Portuguese | 39.8 | 34.9 | 38.9 | 40.2 | 54.5 | 50.9 | 53.3 | 53.4 | 62.1 |
| | | Spanish | 41.5 | 36.0 | 39.8 | 41.8 | 55.1 | 51.9 | 53.6 | 53.6 | 62.5 |
| | | Italian | 39.8 | 34.9 | 38.1 | 40.6 | 53.8 | 49.9 | 52.1 | 51.7 | 61.6 |
| | | German | 39.2 | 34.9 | 37.5 | 39.6 | 53.3 | 50.0 | 52.2 | 51.3 | 60.6 |
| | | French | 40.5 | 34.8 | 39.2 | 40.8 | 54.6 | 51.2 | 53.3 | 53.3 | 62.3 |
| | | Hindi | 33.5 | 30.0 | 32.1 | 34.0 | 43.3 | 40.4 | 42.0 | 42.1 | 50.9 |
| | | Thai | 34.7 | 31.2 | 32.4 | 34.9 | 44.5 | 41.3 | 44.0 | 42.2 | 50.3 |
\*\*for comparison purposes only. Model not released.
## Inference time
In the below table, we compare the performance metrics of different quantization methods (SpinQuant and QAT \+ LoRA) with the BF16 baseline. The evaluation was done using the [ExecuTorch](https://github.com/pytorch/executorch) framework as the inference engine, with the ARM CPU as a backend using Android OnePlus 12 device.
| Category | Decode (tokens/sec) | Time-to-first-token (sec) | Prefill (tokens/sec) | Model size (PTE file size in MB) | Memory size (RSS in MB) |
| :---- | ----- | ----- | ----- | ----- | ----- |
| 1B BF16 (baseline) | 19.2 | 1.0 | 60.3 | 2358 | 3,185 |
| 1B SpinQuant | 50.2 (2.6x) | 0.3 (-76.9%) | 260.5 (4.3x) | 1083 (-54.1%) | 1,921 (-39.7%) |
| 1B QLoRA | 45.8 (2.4x) | 0.3 (-76.0%) | 252.0 (4.2x) | 1127 (-52.2%) | 2,255 (-29.2%) |
| 3B BF16 (baseline) | 7.6 | 3.0 | 21.2 | 6129 | 7,419 |
| 3B SpinQuant | 19.7 (2.6x) | 0.7 (-76.4%) | 89.7 (4.2x) | 2435 (-60.3%) | 3,726 (-49.8%) |
| 3B QLoRA | 18.5 (2.4x) | 0.7 (-76.1%) | 88.8 (4.2x) | 2529 (-58.7%) | 4,060 (-45.3%) |
(\*) The performance measurement is done using an adb binary-based approach.
(\*\*) It is measured on an Android OnePlus 12 device.
(\*\*\*) Time-to-first-token (TTFT) is measured with prompt length=64
*Footnote:*
- *Decode (tokens/second) is for how quickly it keeps generating. Higher is better.*
- *Time-to-first-token (TTFT for shorthand) is for how fast it generates the first token for a given prompt. Lower is better.*
- *Prefill is the inverse of TTFT (aka 1/TTFT) in tokens/second. Higher is better*
- *Model size \- how big is the model, measured by, PTE file, a binary file format for ExecuTorch*
- *RSS size \- Memory usage in resident set size (RSS)*
## Responsibility & Safety
As part of our Responsible release approach, we followed a three-pronged strategy to managing trust & safety risks:
1. Enable developers to deploy helpful, safe and flexible experiences for their target audience and for the use cases supported by Llama
2. Protect developers against adversarial users aiming to exploit Llama capabilities to potentially cause harm
3. Provide protections for the community to help prevent the misuse of our models
### Responsible Deployment
**Approach:** Llama is a foundational technology designed to be used in a variety of use cases. Examples on how Meta’s Llama models have been responsibly deployed can be found in our [Community Stories webpage](https://llama.meta.com/community-stories/). Our approach is to build the most helpful models, enabling the world to benefit from the technology power, by aligning our model safety for generic use cases and addressing a standard set of harms. Developers are then in the driver’s seat to tailor safety for their use cases, defining their own policies and deploying the models with the necessary safeguards in their Llama systems. Llama 3.2 was developed following the best practices outlined in our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/).
#### Llama 3.2 Instruct
**Objective:** Our main objectives for conducting safety fine-tuning are to provide the research community with a valuable resource for studying the robustness of safety fine-tuning, as well as to offer developers a readily available, safe, and powerful model for various applications to reduce the developer workload to deploy safe AI systems. We implemented the same set of safety mitigations as in Llama 3, and you can learn more about these in the Llama 3 [paper](https://ai.meta.com/research/publications/the-llama-3-herd-of-models/).
**Fine-Tuning Data:** We employ a multi-faceted approach to data collection, combining human-generated data from our vendors with synthetic data to mitigate potential safety risks. We’ve developed many large language model (LLM)-based classifiers that enable us to thoughtfully select high-quality prompts and responses, enhancing data quality control.
**Refusals and Tone:** Building on the work we started with Llama 3, we put a great emphasis on model refusals to benign prompts as well as refusal tone. We included both borderline and adversarial prompts in our safety data strategy, and modified our safety data responses to follow tone guidelines.
#### Llama 3.2 Systems
**Safety as a System:** Large language models, including Llama 3.2, **are not designed to be deployed in isolation** but instead should be deployed as part of an overall AI system with additional safety guardrails as required. Developers are expected to deploy system safeguards when building agentic systems. Safeguards are key to achieve the right helpfulness-safety alignment as well as mitigating safety and security risks inherent to the system and any integration of the model or system with external tools. As part of our responsible release approach, we provide the community with [safeguards](https://llama.meta.com/trust-and-safety/) that developers should deploy with Llama models or other LLMs, including Llama Guard, Prompt Guard and Code Shield. All our [reference implementations](https://github.com/meta-llama/llama-agentic-system) demos contain these safeguards by default so developers can benefit from system-level safety out-of-the-box.
### New Capabilities and Use Cases
**Technological Advancement:** Llama releases usually introduce new capabilities that require specific considerations in addition to the best practices that generally apply across all Generative AI use cases. For prior release capabilities also supported by Llama 3.2, see [Llama 3.1 Model Card](https://github.com/meta-llama/llama-models/blob/main/models/llama3_1/MODEL_CARD.md), as the same considerations apply here as well.
**Constrained Environments:** Llama 3.2 1B and 3B models are expected to be deployed in highly constrained environments, such as mobile devices. LLM Systems using smaller models will have a different alignment profile and safety/helpfulness tradeoff than more complex, larger systems. Developers should ensure the safety of their system meets the requirements of their use case. We recommend using lighter system safeguards for such use cases, like Llama Guard 3-1B or its mobile-optimized version.
### Evaluations
**Scaled Evaluations:** We built dedicated, adversarial evaluation datasets and evaluated systems composed of Llama models and Purple Llama safeguards to filter input prompt and output response. It is important to evaluate applications in context, and we recommend building dedicated evaluation dataset for your use case.
**Red Teaming:** We conducted recurring red teaming exercises with the goal of discovering risks via adversarial prompting and we used the learnings to improve our benchmarks and safety tuning datasets. We partnered early with subject-matter experts in critical risk areas to understand the nature of these real-world harms and how such models may lead to unintended harm for society. Based on these conversations, we derived a set of adversarial goals for the red team to attempt to achieve, such as extracting harmful information or reprogramming the model to act in a potentially harmful capacity. The red team consisted of experts in cybersecurity, adversarial machine learning, responsible AI, and integrity in addition to multilingual content specialists with background in integrity issues in specific geographic markets.
### Critical Risks
In addition to our safety work above, we took extra care on measuring and/or mitigating the following critical risk areas:
**1\. CBRNE (Chemical, Biological, Radiological, Nuclear, and Explosive Weapons):** Llama 3.2 1B and 3B models are smaller and less capable derivatives of Llama 3.1. For Llama 3.1 70B and 405B, to assess risks related to proliferation of chemical and biological weapons, we performed uplift testing designed to assess whether use of Llama 3.1 models could meaningfully increase the capabilities of malicious actors to plan or carry out attacks using these types of weapons and have determined that such testing also applies to the smaller 1B and 3B models.
**2\. Child Safety:** Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors including the additional languages Llama 3 is trained on. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.
**3\. Cyber Attacks:** For Llama 3.1 405B, our cyber attack uplift study investigated whether LLMs can enhance human capabilities in hacking tasks, both in terms of skill level and speed.
Our attack automation study focused on evaluating the capabilities of LLMs when used as autonomous agents in cyber offensive operations, specifically in the context of ransomware attacks. This evaluation was distinct from previous studies that considered LLMs as interactive assistants. The primary objective was to assess whether these models could effectively function as independent agents in executing complex cyber-attacks without human intervention. Because Llama 3.2’s 1B and 3B models are smaller and less capable models than Llama 3.1 405B, we broadly believe that the testing conducted for the 405B model also applies to Llama 3.2 models.
### Community
**Industry Partnerships:** Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership on AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama).
**Grants:** We also set up the [Llama Impact Grants](https://llama.meta.com/llama-impact-grants/) program to identify and support the most compelling applications of Meta’s Llama model for societal benefit across three categories: education, climate and open innovation. The 20 finalists from the hundreds of applications can be found [here](https://llama.meta.com/llama-impact-grants/#finalists).
**Reporting:** Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community.
## Ethical Considerations and Limitations
**Values:** The core values of Llama 3.2 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3.2 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.
**Testing:** Llama 3.2 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3.2’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3.2 models, developers should perform safety testing and tuning tailored to their specific applications of the model. Please refer to available resources including our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide), [Trust and Safety](https://llama.meta.com/trust-and-safety/) solutions, and other [resources](https://llama.meta.com/docs/get-started/) to learn more about responsible development.
|
dimasik87/ff746235-8f40-440d-bbac-a0bad1a05d33 | dimasik87 | 2025-05-29T12:47:08Z | 0 | 0 | peft | [
"peft",
"safetensors",
"mistral",
"axolotl",
"generated_from_trainer",
"base_model:unsloth/Mistral-Nemo-Base-2407",
"base_model:adapter:unsloth/Mistral-Nemo-Base-2407",
"license:apache-2.0",
"4-bit",
"bitsandbytes",
"region:us"
] | null | 2025-05-29T09:01:49Z | ---
library_name: peft
license: apache-2.0
base_model: unsloth/Mistral-Nemo-Base-2407
tags:
- axolotl
- generated_from_trainer
model-index:
- name: ff746235-8f40-440d-bbac-a0bad1a05d33
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
<details><summary>See axolotl config</summary>
axolotl version: `0.4.1`
```yaml
absolute_data_files: false
adapter: lora
base_model: unsloth/Mistral-Nemo-Base-2407
bf16: true
chat_template: llama3
dataset_prepared_path: /workspace/axolotl
datasets:
- data_files:
- b737943c169dce76_train_data.json
ds_type: json
format: custom
path: /workspace/input_data/
type:
field_input: input
field_instruction: instruct
field_output: output
format: '{instruction} {input}'
no_input_format: '{instruction}'
system_format: '{system}'
system_prompt: ''
debug: null
deepspeed: null
dpo:
beta: 0.1
enabled: true
group_by_length: false
rank_loss: true
reference_model: null
early_stopping_patience: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 1
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
gradient_clipping: 1.0
group_by_length: false
hub_model_id: dimasik87/ff746235-8f40-440d-bbac-a0bad1a05d33
hub_repo: null
hub_strategy: end
hub_token: null
learning_rate: 1.0e-06
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 64
lora_dropout: 0.1
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lr_scheduler: cosine
max_steps: 500
micro_batch_size: 6
mixed_precision: bf16
mlflow_experiment_name: /tmp/b737943c169dce76_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 2
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
saves_per_epoch: 1
sequence_len: 1024
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 9f4dd85f-eef8-4321-bcfe-a15029c10fe9
wandb_project: s56-7
wandb_run: your_name
wandb_runid: 9f4dd85f-eef8-4321-bcfe-a15029c10fe9
warmup_steps: 50
weight_decay: 0.05
xformers_attention: true
```
</details><br>
# ff746235-8f40-440d-bbac-a0bad1a05d33
This model is a fine-tuned version of [unsloth/Mistral-Nemo-Base-2407](https://huggingface.co/unsloth/Mistral-Nemo-Base-2407) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4468
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-06
- train_batch_size: 6
- eval_batch_size: 6
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 24
- optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 50
- training_steps: 500
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 6.5079 | 0.0000 | 1 | 1.6777 |
| 6.0877 | 0.0099 | 250 | 1.4972 |
| 5.6748 | 0.0198 | 500 | 1.4468 |
### Framework versions
- PEFT 0.13.2
- Transformers 4.46.0
- Pytorch 2.5.0+cu124
- Datasets 3.0.1
- Tokenizers 0.20.1 |
sergioalves/0c81828f-da57-4c8e-8422-1f63b7f3acff | sergioalves | 2025-05-29T12:46:23Z | 0 | 0 | peft | [
"peft",
"safetensors",
"mistral",
"axolotl",
"generated_from_trainer",
"base_model:unsloth/Mistral-Nemo-Base-2407",
"base_model:adapter:unsloth/Mistral-Nemo-Base-2407",
"license:apache-2.0",
"4-bit",
"bitsandbytes",
"region:us"
] | null | 2025-05-29T09:01:49Z | ---
library_name: peft
license: apache-2.0
base_model: unsloth/Mistral-Nemo-Base-2407
tags:
- axolotl
- generated_from_trainer
model-index:
- name: 0c81828f-da57-4c8e-8422-1f63b7f3acff
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
<details><summary>See axolotl config</summary>
axolotl version: `0.4.1`
```yaml
absolute_data_files: false
adapter: lora
base_model: unsloth/Mistral-Nemo-Base-2407
bf16: true
chat_template: llama3
dataset_prepared_path: /workspace/axolotl
datasets:
- data_files:
- b737943c169dce76_train_data.json
ds_type: json
format: custom
path: /workspace/input_data/
type:
field_input: input
field_instruction: instruct
field_output: output
format: '{instruction} {input}'
no_input_format: '{instruction}'
system_format: '{system}'
system_prompt: ''
debug: null
deepspeed: null
dpo:
beta: 0.1
enabled: true
group_by_length: false
rank_loss: true
reference_model: null
early_stopping_patience: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 1
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
gradient_clipping: 0.85
group_by_length: false
hub_model_id: sergioalves/0c81828f-da57-4c8e-8422-1f63b7f3acff
hub_repo: null
hub_strategy: end
hub_token: null
learning_rate: 1.0e-06
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 64
lora_dropout: 0.1
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lr_scheduler: cosine
max_steps: 500
micro_batch_size: 6
mixed_precision: bf16
mlflow_experiment_name: /tmp/b737943c169dce76_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 2
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
saves_per_epoch: 1
sequence_len: 1024
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 9f4dd85f-eef8-4321-bcfe-a15029c10fe9
wandb_project: s56-7
wandb_run: your_name
wandb_runid: 9f4dd85f-eef8-4321-bcfe-a15029c10fe9
warmup_steps: 50
weight_decay: 0.05
xformers_attention: true
```
</details><br>
# 0c81828f-da57-4c8e-8422-1f63b7f3acff
This model is a fine-tuned version of [unsloth/Mistral-Nemo-Base-2407](https://huggingface.co/unsloth/Mistral-Nemo-Base-2407) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4497
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-06
- train_batch_size: 6
- eval_batch_size: 6
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 24
- optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 50
- training_steps: 500
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 6.5079 | 0.0000 | 1 | 1.6777 |
| 6.1138 | 0.0099 | 250 | 1.5015 |
| 5.6909 | 0.0198 | 500 | 1.4497 |
### Framework versions
- PEFT 0.13.2
- Transformers 4.46.0
- Pytorch 2.5.0+cu124
- Datasets 3.0.1
- Tokenizers 0.20.1 |
MasterKoco/SoundsRight_DENOISING_16000HZ | MasterKoco | 2025-05-29T12:42:33Z | 0 | 0 | null | [
"region:us"
] | null | 2025-05-26T11:46:37Z | # Container Template for SoundsRight Subnet Miners
This repository contains a contanierized version of [SGMSE+](https://huggingface.co/sp-uhh/speech-enhancement-sgmse) and serves as a tutorial for miners to format their models on [Bittensor's](https://bittensor.com/) [SoundsRight Subnet](https://github.com/synapsec-ai/SoundsRightSubnet). The branches `DENOISING_16000HZ` and `DEREVERBERATION_16000HZ` contain SGMSE fitted with the approrpriate checkpoints for denoising and dereverberation tasks at 16kHz, respectively.
This container has only been tested with **Ubuntu 24.04** and **CUDA 12.6**. It may run on other configurations, but it is not guaranteed.
To run the container, first configure NVIDIA Container Toolkit and generate a CDI specification. Follow the instructions to download the [NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html) with Apt.
Next, follow the instructions for [generating a CDI specification](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/cdi-support.html).
Verify that the CDI specification was done correctly with:
```
$ nvidia-ctk cdi list
```
You should see this in your output:
```
nvidia.com/gpu=all
nvidia.com/gpu=0
```
If you are running podman as root, run the following command to start the container:
Run the container with:
```
podman build -t modelapi . && podman run -d --device nvidia.com/gpu=all --user root --name modelapi -p 6500:6500 modelapi
```
Access logs with:
```
podman logs -f modelapi
```
If you are running the container rootless, there are a few more changes to make:
First, modify `/etc/nvidia-container-runtime/config.toml` and set the following parameters:
```
[nvidia-container-cli]
no-cgroups = true
[nvidia-container-runtime]
debug = "/tmp/nvidia-container-runtime.log"
```
You can also run the following command to achieve the same result:
```
$ sudo nvidia-ctk config --set nvidia-container-cli.no-cgroups --in-place
```
Run the container with:
```
podman build -t modelapi . && podman run -d --device nvidia.com/gpu=all --volume /usr/local/cuda-12.6:/usr/local/cuda-12.6 --user 10002:10002 --name modelapi -p 6500:6500 modelapi
```
Access logs with:
```
podman logs -f modelapi
```
Running the container will spin up an API with the following endpoints:
1. `/status/` : Communicates API status
2. `/prepare/` : Download model checkpoint and initialize model
3. `/upload-audio/` : Upload audio files, save to noisy audio directory
4. `/enhance/` : Initialize model, enhance audio files, save to enhanced audio directory
5. `/download-enhanced/` : Download enhanced audio files
By default the API will use host `0.0.0.0` and port `6500`.
### References
1. **Welker, Simon; Richter, Julius; Gerkmann, Timo**
*Speech Enhancement with Score-Based Generative Models in the Complex STFT Domain*.
Proceedings of *Interspeech 2022*, 2022, pp. 2928–2932.
[DOI: 10.21437/Interspeech.2022-10653](https://doi.org/10.21437/Interspeech.2022-10653)
2. **Richter, Julius; Welker, Simon; Lemercier, Jean-Marie; Lay, Bunlong; Gerkmann, Timo**
*Speech Enhancement and Dereverberation with Diffusion-based Generative Models*.
*IEEE/ACM Transactions on Audio, Speech, and Language Processing*, Vol. 31, 2023, pp. 2351–2364.
[DOI: 10.1109/TASLP.2023.3285241](https://doi.org/10.1109/TASLP.2023.3285241)
3. **Richter, Julius; Wu, Yi-Chiao; Krenn, Steven; Welker, Simon; Lay, Bunlong; Watanabe, Shinjii; Richard, Alexander; Gerkmann, Timo**
*EARS: An Anechoic Fullband Speech Dataset Benchmarked for Speech Enhancement and Dereverberation*.
Proceedings of *ISCA Interspeech*, 2024, pp. 4873–4877.
|
JesseLiu/llama32-3b-pagerank-partial-naive-grpo-lora | JesseLiu | 2025-05-29T12:33:21Z | 0 | 0 | peft | [
"peft",
"safetensors",
"arxiv:1910.09700",
"base_model:meta-llama/Llama-3.2-3B-Instruct",
"base_model:adapter:meta-llama/Llama-3.2-3B-Instruct",
"region:us"
] | null | 2025-05-29T12:32:45Z | ---
base_model: meta-llama/Llama-3.2-3B-Instruct
library_name: peft
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
### Framework versions
- PEFT 0.15.1 |
JTh34/puppy-embed-colab-d23f57a4 | JTh34 | 2025-05-29T12:31:00Z | 0 | 0 | sentence-transformers | [
"sentence-transformers",
"safetensors",
"bert",
"sentence-similarity",
"feature-extraction",
"generated_from_trainer",
"dataset_size:700",
"loss:MatryoshkaLoss",
"loss:MultipleNegativesRankingLoss",
"arxiv:1908.10084",
"arxiv:2205.13147",
"arxiv:1705.00652",
"base_model:Snowflake/snowflake-arctic-embed-l",
"base_model:finetune:Snowflake/snowflake-arctic-embed-l",
"model-index",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | sentence-similarity | 2025-05-29T12:29:55Z | ---
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- generated_from_trainer
- dataset_size:700
- loss:MatryoshkaLoss
- loss:MultipleNegativesRankingLoss
base_model: Snowflake/snowflake-arctic-embed-l
widget:
- source_sentence: How can I ensure that my puppy's emotional and psychological needs
are being met during training to promote better social skills and emotional stability?
sentences:
- Look for wiggling body language, happy faces, and play bows from both dogs. Bouncy
exaggerated rocking-horse type movements are a sign the dogs are having fun. Determine
whether your dog wants to run back and play with another dog by giving a consent
test. Separate the dogs, remove your dog some distance away, then observe whether
your dog wants to run back and play with the other dog or not. Dogs should be
self-imposing breaks from play to rest or get a small drink. Breaks may include
sniffing. If one dog wants a break the other dog shows respect by not persisting
in perpetual play inducements. If your dog comes and lies down near you, take
the lead, and protect your dog from further interaction until your dog desires
another round of play. Sniffing may be in order. Puppies and adult dogs must be
monitored very carefully as should small dogs vs. large dogs. Not all adult dogs
like puppies. Large adult dogs should change levels, that is voluntarily self-
handicap, and let the puppy or smaller dog “win” regularly during role reversals.
These behaviors balance inequalities in size, strength, and health. If role-reversals
or self-handicapping are not occurring, do not allow your puppy or small dog continued
interaction with an adult or much larger dog who may show aggression or cause
physical or emotional injury. One dog may be either the victim or the bully in
different contexts. Roles may also change when playing with different play partners.
Dogs should take turns happily chasing each other where neither dog is a bully
or a target, so play goes back and forth. With puppies, adult dogs should be willing
to let the puppy win now and again and should always back off if the puppy squeals.
- 'The priorities, best practices and the exclusions define what force-free training
is and in addition what it is not. Force-free training is categorically and qualitatively
different and unique from traditional or so-called “balance training” by virtue
of the operational definition of what we do and what we do not do as described
below.
No shock No prong No choke No pain No fear No dominance or intimidation No compulsion
methods No physical force No hitting with any object, including rolled up towels
No throwing items No swatting with a newspaper No shaking cans of coins or rocks
in our dogs’ sensitive ears No spraying water in the face or body No yelling
Nurturing biological health and psychological well-being enhances emotional stability,
social skills, and cognitive abilities by meeting our dogs’ real needs. The Hierarchy
of Dog Needs describes the emotional and behavioral modification methods that force-free
behavior modification consultants and trainers endorse. Effectively using these
techniques serves to increase, decrease, and redirect behavior, and also to change
emotional responses. We set the stage for optimal well-being by using force-free
training, and eliminating the potentiality for fear, stress, and aggression.'
- 'Many properties of commercially prepared dog food are not sufficiently regulated.
We suggest that you
Read the ingredients on labels and choose a food with a specifically named protein
source as the first ingredient. Avoid the vague term “meat,” by-products, corn
syrup, and sugar. Meat meal is generally rendered meat made of by-products and
that is why you will not see meat meals in human foods. Avoid meat meals when
there are higher quality alternatives. Avoid artificial flavors, colors, and preservatives,
especially BHA, BHT, and ethoxyquin that may be linked to carcinogens. Artificial
preservatives may be toxic to your dogs: Some artificial preservatives are also
used as pesticides.
Consider the following questions when choosing a brand of dog food
Sources. Where do the ingredients come from? Recalls. What is the manufacturer’s
safety record? Marketing and customer service. Is the company transparent about
its products and responsive to inquiries?'
- source_sentence: How can I ensure my puppy doesn't have potty accidents when I'm
not watching her closely?
sentences:
- 'attached to you by a leash so she canʼt wander off to potty in the house
or under your direct supervision in an enclosed area. Direct supervision means
you are looking at her at all times. The minute you turn away, sheʼll have a potty
accident.
Stick to this plan for a month straight and sheʼll reliably develop the habit
of going outside and holding it inside. Then continue keeping a close eye on her
for another couple of months, especially when you take her on outings to other
peopleʼs homes, before declaring her completely potty trained.
The goal of crate training is that your puppy learns to love resting in her crate.
Crate: Your puppy should sleep in her crate at night and take naps in it during
the day. To train her to love her crate, you can make it comfortable with a blanket
and
place treats inside at random times. Then give her toys and pet her when sheʼs
in it
before you close the door. The ultimate goal of crate training is that she goes
into
the crate on her own or when you give her a verbal cue, rather than needing to
be
shoved or coaxed in. And once sheʼs in, she remains calm, relaxed and quiet. (If
you have problems with this, download the crate training handout at
Crate size: The crate should be big enough for the puppy to lie down and turn
around but not big enough for a separate potty area. You can make the crate
smaller by placing a box in it and, as the puppy grows, enlarge the crate by using
a
“The goal of crate training is that your puppy learns to love resting in her crate.”'
- 'Does your puppy love their food? If so, use it to reward them throughout the
day. Portion out some or all of the food and use it to motivate quick responses
and self-control. (For more on how to use food to inspire learning, check out
Part 4 of this book.)
Water is critically important for your puppy’s well-being: it should be left out
and available at all times. That said, try to monitor their drinking habits while
house- training them. Establish a drinking station for your puppy and keep their
dish there, whether it’s empty or full. Give water with meals, after playing,
chewing, or napping, and as you’re on your way to the potty area.
Restrict water after 7:30 p.m., unless you want to be up all night taking your
puppy outside. If your puppy needs a drink, either give them a small amount or
offer a couple of ice cubes.
Although dogs have many fewer taste buds overall (humans have 9,000 to their 1,700),
your puppy has a ring of taste buds on the tip of their tongue that make water
taste sweet. Pretty cool.
I don’t think house-training can be summed up any better than with the wonder-
ful maxim “Whatever goes in must come out.” Your puppy’s biological clock will
have them eliminating on demand. When their bladder or bowels are pressed, they’ll
let loose whether they’re outside or on the papers — or the rug, if you’re not
watching.'
- 'After a puppy’s peak socialization period, around 16 weeks of age, it’s impos-
sible to turn back the clock. People, places, sights, and smells that your puppy
would have conditioned to naturally at an early age will seem suspicious to an
older puppy. Do you want a dog who can’t warm up to everyday stimulations?
Puppies who are overisolated or stressed during infancy are shown to chew more
destructively and may wreck your furnishings if they aren’t conditioned to chew
their toys. The early turmoil created nervous energy that needs to be displaced,
and because running to the refrigerator is off limits and nail-biting
isn’t an option, your puppy will chew on whatever is available. Provide plenty
of satisfying options or else you may see your sofa disappear, one cushion at
a time.
» Shelter: If you find an older puppy at a shelter, ask about their history and
try to find out why they were left there. Were they found on the side of the street,
or have they grown up in the system? Has the puppy in question been returned more
than once? Ask what the reasons were — you may be adopting a dog who couldn’t
be house-trained, was fearful of kids, or showed aggression when chewing a bone.
Find out what the staff thinks of the puppy’s personality. » Pet store, puppy
broker, craigslist, and other sources: Discount shopping isn’t for puppies. Do
not buy a puppy without meeting and talking to the breeder or rescuers first.
Though you may read a phrase that makes you feel like you’ll be the winner, there
are no winners in the online puppy shopping game. Buying a puppy in this manner
ensures that more puppies will be bred this way, which doesn’t take their interests
to heart.'
- source_sentence: What steps should I follow to transition my puppy from walking
in a low-distraction area to more crowded places while using a Halti?
sentences:
- examine teeth and inside the mouth. This will desensitize your dog to people touching
your dog’s mouth. Make sure your dog has proper dental care. Proper dental care
is so often overlooked! Massage the sacral joint where the spine meets the tail
and lift the tail to desensitize your dog to handling. Teach your dog how to give
consent, how to ask for a break, and how to ask you to stop. Choice and consent
are the wave of the future in modern cooperative care.
- 'If your adolescent puppy is still pulling on the leash despite all the basic
leash walking training I’ve taken you through, you might choose to try a Halti
leash. I’m a big fan of Haltis, as they make a massive difference with dogs who
continue to have pulling problems, especially reactive dogs who can pull unexpectedly.
A Halti is a gentle head harness that controls the dog by the snout and takes
all the pressure off the throat.
To introduce a Halti, you must first make sure your puppy is comfortable putting
their nose into the Halti. Do this by simply slipping it on and off and rewarding
your puppy heavily for this (YES, and a treat). You’ll then need to get your puppy
comfortable with having the Halti done up, so again, go heavy on the rewards and
praise. Then you need to master walking your puppy with a Halti on in a low-distraction
environment (i.e., your garden or somewhere else familiar, without loads of other
dogs and people). Once your dog is comfortable moving around your garden, you
can move to outside; get your dog comfortable in your street before moving on
to high- distraction-level areas, such as parks. A decent structured walk should
be ten minutes, with your dog nicely walking by your side, followed by five minutes
where you allow your puppy to sniff and explore a little, before returning to
a more structured HEEL walk.
TIP: Haltis can be tricky to put on at first, so watch a few online videos or ask
a friend
to demonstrate theirs just to get a feel for how they work. Waving a Halti around
and'
- 'Think of your dog’s veterinarian as being on par with your doctor or your child’s
pediatrician. Medical knowledge is essential, but a good bedside manner is the
cherry on top of the sundae. Speak with the receptionists and bring in your pup
for a cheerful social call before their initial visit. Talk to the doctor like
they’re a neighbor. Do you feel comfortable sharing all your canine concerns with
them?
If you’re unsure of which veterinarian to use, ask around. You can narrow your
search by asking your friends and family whom they use and why.
Puppies can be quite impulsive — they often swallow things that look edible before
even considering whether they are. So, at your puppy’s first veterinary visit,
ask the doctor for a recommended method for inducing vomiting. You should also
find out the poison-control hotline number and always keep it on your phone in
case of an emergency. Seek out a 24-hour emergency veterinary hospital in your
area as well. Keep the hospital’s number by or on your phone. Accidents can happen
during off hours, so have a plan.
Whether your life demands consistent hours away from home or circumstance steps
in to temporarily rearrange your schedule, knowing a dog walker can make the difference
between a happy puppy and a stressed-out one. Puppies are like human babies in
that they have a strong need dependency. Even though an adult dog can hold their
bladder until you get home or can survive until a late meal, your puppy may well
eat the walls of your house if you get stuck in traffic. A reliable dog walker
can be a godsend in times like these.'
- source_sentence: What breeds of dogs might struggle more with swimming, and how
can I support them if they are hesitant to enter the water?
sentences:
- 'Never throw your dog into the water to “see how it goes”—because your dog can
easily drown, struggle or thrash and any trust you may have established will be
broken. The safest, least stressful, and most effective way to teach a dog to
swim is to use a properly-fitted life jacket. A life jacket often helps new swimmers
relax enough to paddle with all four legs.
Desensitize your dog to wearing a dog life jacket to keep him afloat and to provide
peace of mind for you. Your dog’s innate ability to swim or ease in learning to
swim is, in part, determined by breed and body morphology. Even some retrievers
need a helping hand to learn and find the confidence to swim for fun and exercise.
Breeds with short legs and wide chests such as Bulldogs, Boston Terriers, Corgis,
and Pugs, simply were not bred for swimming. Large, muscled breeds such as the
bully breeds, require a great deal of energy expenditure in the water due to their
significant body mass. Sight hounds, such as Salukis, Whippets, Italian Greyhounds
and Greyhounds, have the disadvantage of both large muscles and little body fat
to keep them afloat.'
- 'The key to etiquette training is to set your goals and share them with family
and friends — and even with strangers who interact with your pup. Think of this
last training chapter of Part 3 as sending your puppy off to Miss Sarah’s School
of Dog Etiquette, which is a short-term course with long-term freedoms and rewards.
To develop the all-important canine consciousness, you must do two things:
» Decide what you want when you give a direction. » Follow through — if your expectations
are unclear, your puppy’s reaction will
When debuting that almost-grown puppy of yours, follow these five essential rules:
1. Make sure your puppy is familiar and comfortable with the setting before you
attempt to introduce them to anyone. Don’t greet people your first day out.
2. Before each introduction, insist that your puppy stand still at your side.
Gently hold still or bring them back to your side and instruct “Wait.”
3. Tell admirers “We’re in training.” This statement will help them respect your
efforts and contain their excitement (hopefully).
4. Stay more focused on your puppy than on the admirer. Insist that your puppy
use good manners before you let them approach a new dog or person. 5. Put faith
in your knowledge. Just because everyone has advice doesn’t mean they’re right.
“I don’t mind if they jump” doesn’t hold water. You mind if they jump, so don’t
give in.
Under and back: Helpful commands when you’re out and about
Have you ever marveled at the sight of a dog lying patiently under the table or
their human’s legs? It’s calming on all fronts because the dog is at peace knowing
that the person is safe and in charge. Fortunately for you, it couldn’t be easier
to teach your pup this skill.'
- 'exposed to an aversive stimulus that it cannot escape or avoid, and which nothing
it does has any effect on, eventually its avoidance responses will extinguish.
It will stop reacting to the stimulus, pay no attention, and apparently become
unaware of it. This is called habituation. In my New York apartment I found the
street noise unbearable at first, but eventually, like most New Yorkers, I learned
to sleep through the sirens, yelling, garbage trucks, even car crashes. I became
habituated. Police horses are sometimes trained by subjecting them to all kinds
of harmless but alarming events, such as opening umbrellas, flapping papers, being
tapped all over with rattling tin cans, and so on. The horses become so habituated
to startling sights and sounds that they remain unflappable no matter what events
the city streets have to offer.
Method 4 is not useful for getting rid of well-learned, self-rewarding behavior
patterns. It is good, however, for whining, sulking, or teasing. Even small children
can learn - and are gratified to discover - that they can stop older children from
teasing them merely by not reacting in any way, good or bad.
BEHAVIOR Roommate leaves dirty laundry all over the place. Dog in yard barks all
night.
This behavior is self-reinforcing and seldom extinguishes spontaneously. A certain
amount of noise is natural and harmless; let it be, they''ll get tired of it.
See to it that his or her harsh words have no results, either good or bad. Work
on other strokes, footwork, and so on, and try to let the specific error die down
from lack of concentrating on it. If the misbehavior is a way of getting attention,
remove the attention; shirking,'
- source_sentence: What should I do if I notice signs of an ear infection in my puppy?
sentences:
- 'Puppies love interactive games, especially as they mature. Tug is a great game
for puppies and can be used to teach your puppy to Give on command, as well as
learning what is and isn’t okay to tug on — your hair or slippers, for example.
Here’s a quick lesson on playing tug with a young puppy:
» Start with a rope or doggie play pole, which can also be fashioned out of a
» Bounce the toy in front of your puppy or wait until they show interest in
playing with it. Reward their interest by saying “Tug” and providing resistance.
» Take a smelly food treat in your hand (like liver, hot dog, or jerky-type treat)
» Let your puppy have the toy back right away and continue playing or say “You
Soon your puppy will learn that sharing and releasing toys means more fun and
interaction, not less.
When picking out self-soothing toys for your puppy (objects they can play with
alone), keep the analogy of giving a child your smartphone to keep them busy when
you’re present but not accessible. Self-soothing objects come in many forms: What
calms your puppy best?
Though you generally can’t go wrong with indestructible plastic bones, some puppies
find them, well, boring. Rawhide, which is made in America, is a satisfy- ing
chew, but it’s problematic with some dogs who chew obsessively because they gulp
it as they go and can choke or get indigestion. Destructible bones also cost money
to replace — just saying.
Personally, my clients have had the most luck with pressed rawhide, animal-part
sticks (hooves and bully sticks), and vegetable-matter pulp bones. Test out a
few kinds yourself to find a bone that satisfies your puppy’s craving and that
can pass the “systems” test (their digestive system, that is); then buy it in
bulk.'
- 'Don’t use cotton swabs or poke into your puppy’s ear canal. You can cause irreparable
damage by doing so.
» Prevent water from entering the ear. If you’re bathing your pup, put a cotton
ball in the opening ahead of time and wipe the ears out with a dry cotton ball
when you’re finished.
Ear infections are quite common. Signs of infection include a red or swollen ear,
discharge, head shaking, ear itching, or bad odor. If you notice any of these
symp- toms, get your puppy to their doctor immediately. Left untreated, infections
can cause fever, depression, irritability, and loss of balance. Your veterinarian
may prescribe an ointment that you administer at home. Here’s how to use it:
1. Wait until your dog’s a little sleepy. 2. Bring them to the refrigerator and
swipe some peanut butter or broth at their eye level.
3. As they’re licking the refrigerator, gently squeeze into their ear canal the
amount of ointment specified by your veterinarian.
You don’t have to know much about the nose, though it is helpful for tipping you
off to the fact that your puppy’s not feeling well. A warm nose can be caused
by elevated temperature. (See the nearby sidebar, “Taking your puppy’s tempera-
ture.”) However, weather conditions also can lead to dryness or fluctuation in
body temperature. If you suspect that your puppy has a fever, touch their other
body areas without fur (belly, paws, or the inside of their ears) or take their
tem- perature. Did I mention that you have to do it rectally? What fun!'
- 'More than 320 breeds are now registered worldwide. These days, being a purebred
dog is like belonging to an exclusive club: Only dogs with similar looks and inter-
ests get in. Although most breeds are no longer asked to do the work they were
developed for, fanciers continually devote themselves to breeding and selling
puppies that reflect their traditions.
Choosing a specific breed enables you to predict the size, weight, and interest
of your puppy. Selecting a one-of-a-kind mixed-breed puppy, and predicting or
discovering the various breeds that combine to create them, allows you to make
accurate descriptions about their interests and energy level as an adult dog.
When researching a breed, mixed-breed, or designer-mixed-breed, try to meet at
least three adult dogs of the same breed or mix-breeds. All puppies are cute and
adorable, but they grow up in the blink of an eye, so make sure you like the look
and personality of the dog your puppy will become.
Whether you’re considering a purebred, mixed-breed, or designer-mixed-breed, take
a good, hard look at your lifestyle now and project out five to ten years. How
might a certain breed’s or mixed breed’s interests and energy level play out in
your home?'
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
- cosine_accuracy@1
- cosine_accuracy@3
- cosine_accuracy@5
- cosine_accuracy@10
- cosine_precision@1
- cosine_precision@3
- cosine_precision@5
- cosine_precision@10
- cosine_recall@1
- cosine_recall@3
- cosine_recall@5
- cosine_recall@10
- cosine_ndcg@10
- cosine_mrr@10
- cosine_map@100
model-index:
- name: SentenceTransformer based on Snowflake/snowflake-arctic-embed-l
results:
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: Unknown
type: unknown
metrics:
- type: cosine_accuracy@1
value: 0.6666666666666666
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.8533333333333334
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.9266666666666666
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.96
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.6666666666666666
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.2844444444444444
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.1853333333333333
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.09599999999999996
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.6666666666666666
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.8533333333333334
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.9266666666666666
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.96
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.8201527661146794
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.7745740740740741
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.7766199861997735
name: Cosine Map@100
---
# SentenceTransformer based on Snowflake/snowflake-arctic-embed-l
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-l](https://huggingface.co/Snowflake/snowflake-arctic-embed-l). It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
## Model Details
### Model Description
- **Model Type:** Sentence Transformer
- **Base model:** [Snowflake/snowflake-arctic-embed-l](https://huggingface.co/Snowflake/snowflake-arctic-embed-l) <!-- at revision d8fb21ca8d905d2832ee8b96c894d3298964346b -->
- **Maximum Sequence Length:** 512 tokens
- **Output Dimensionality:** 1024 dimensions
- **Similarity Function:** Cosine Similarity
<!-- - **Training Dataset:** Unknown -->
<!-- - **Language:** Unknown -->
<!-- - **License:** Unknown -->
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): Normalize()
)
```
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("JTh34/puppy-embed-colab-d23f57a4")
# Run inference
sentences = [
'What should I do if I notice signs of an ear infection in my puppy?',
'Don’t use cotton swabs or poke into your puppy’s ear canal. You can cause irreparable damage by doing so.\n\n» Prevent water from entering the ear. If you’re bathing your pup, put a cotton ball in the opening ahead of time and wipe the ears out with a dry cotton ball when you’re finished.\n\nEar infections are quite common. Signs of infection include a red or swollen ear, discharge, head shaking, ear itching, or bad odor. If you notice any of these symp- toms, get your puppy to their doctor immediately. Left untreated, infections can cause fever, depression, irritability, and loss of balance. Your veterinarian may prescribe an ointment that you administer at home. Here’s how to use it:\n\n1. Wait until your dog’s a little sleepy. 2. Bring them to the refrigerator and swipe some peanut butter or broth at their eye level.\n\n3. As they’re licking the refrigerator, gently squeeze into their ear canal the amount of ointment specified by your veterinarian.\n\nYou don’t have to know much about the nose, though it is helpful for tipping you off to the fact that your puppy’s not feeling well. A warm nose can be caused by elevated temperature. (See the nearby sidebar, “Taking your puppy’s tempera- ture.”) However, weather conditions also can lead to dryness or fluctuation in body temperature. If you suspect that your puppy has a fever, touch their other body areas without fur (belly, paws, or the inside of their ears) or take their tem- perature. Did I mention that you have to do it rectally? What fun!',
'More than 320 breeds are now registered worldwide. These days, being a purebred dog is like belonging to an exclusive club: Only dogs with similar looks and inter- ests get in. Although most breeds are no longer asked to do the work they were developed for, fanciers continually devote themselves to breeding and selling puppies that reflect their traditions.\n\nChoosing a specific breed enables you to predict the size, weight, and interest of your puppy. Selecting a one-of-a-kind mixed-breed puppy, and predicting or discovering the various breeds that combine to create them, allows you to make accurate descriptions about their interests and energy level as an adult dog.\n\nWhen researching a breed, mixed-breed, or designer-mixed-breed, try to meet at least three adult dogs of the same breed or mix-breeds. All puppies are cute and adorable, but they grow up in the blink of an eye, so make sure you like the look and personality of the dog your puppy will become.\n\nWhether you’re considering a purebred, mixed-breed, or designer-mixed-breed, take a good, hard look at your lifestyle now and project out five to ten years. How might a certain breed’s or mixed breed’s interests and energy level play out in your home?',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
```
<!--
### Direct Usage (Transformers)
<details><summary>Click to see the direct usage in Transformers</summary>
</details>
-->
<!--
### Downstream Usage (Sentence Transformers)
You can finetune this model on your own dataset.
<details><summary>Click to expand</summary>
</details>
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
## Evaluation
### Metrics
#### Information Retrieval
* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.6667 |
| cosine_accuracy@3 | 0.8533 |
| cosine_accuracy@5 | 0.9267 |
| cosine_accuracy@10 | 0.96 |
| cosine_precision@1 | 0.6667 |
| cosine_precision@3 | 0.2844 |
| cosine_precision@5 | 0.1853 |
| cosine_precision@10 | 0.096 |
| cosine_recall@1 | 0.6667 |
| cosine_recall@3 | 0.8533 |
| cosine_recall@5 | 0.9267 |
| cosine_recall@10 | 0.96 |
| **cosine_ndcg@10** | **0.8202** |
| cosine_mrr@10 | 0.7746 |
| cosine_map@100 | 0.7766 |
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Dataset
#### Unnamed Dataset
* Size: 700 training samples
* Columns: <code>sentence_0</code> and <code>sentence_1</code>
* Approximate statistics based on the first 700 samples:
| | sentence_0 | sentence_1 |
|:--------|:-----------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|
| type | string | string |
| details | <ul><li>min: 15 tokens</li><li>mean: 23.47 tokens</li><li>max: 40 tokens</li></ul> | <ul><li>min: 44 tokens</li><li>mean: 331.62 tokens</li><li>max: 512 tokens</li></ul> |
* Samples:
| sentence_0 | sentence_1 |
|:--------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| <code>What techniques can I apply to train my puppy to stay calm and still when I approach, similar to how llamas are trained?</code> | <code>are each a process, defined by results. Negative reinforcers can be used effectively to train behavior, and even though aversive stimuli are involved, the process can be relatively benign. Here (with thanks to llama expert Jim Logan) is a nice use of the negative reinforcer with a semidomestic animal, the llama, kept in the United States as pets and elsewhere as pack animals and for their wool.<br><br>Llamas are timid and shy, like horses. Unless handled a lot when young,<br><br>they can be hard to approach. So, while operant conditioning with a food reinforcer works splendidly with llamas, if a llama is too skittish to come close enough to a person to take the food, here's what modern llama trainers do. They use a clicker as a signal to tell the llama that what it is doing has earned a reinforcer, but the primary or real reinforcer is the removal of a negative reinforcer, an aversive.<br><br>In effect, you say to the llama, "Will you stand still if I approach within<br><br>thirty feet? Yes? Good. I'll click m...</code> |
| <code>What are the best ways to socialize my hound puppy with household pets to avoid any chasing instincts?</code> | <code>When these puppies are exercised, directed, and included, no group is more happy-go-lucky and accepting of life’s random chaos. But when they don’t get enough playtime or training, they can be hyperactive and destructive.<br><br>Even though the loyal and cheerful dogs in the Sporting group have well-earned reputations as patient family pets, they need both mental and physical stimulation. They can’t cope with long hours of isolation; coupled with a lack of exercise, this isolation fuels anxiety. An unhappy Sporting dog is destructive, hyperactive, and impulsive. This isn’t a good mix — especially for your couch and end table.<br><br>The dogs in the Hound group are a happy lot with a 1-track mind; their fascination with hunting propels them through life and allows them plenty of opportunity for employment. Though you may have no interest in hunting a fox, chasing deer, or treeing a raccoon, your hound puppy probably will.<br><br>Originally teamed in pairs or packs, each hound was prized for their instinc...</code> |
| <code>What are the top five dog breeds recommended for first-time owners, and what makes them suitable for beginners?</code> | <code>Now don’t get me wrong, I’m not trying to put you off. I love dogs, and I think everyone can benefit from having one in their life. If you’re still unsure which breed is right for you, let me suggest a few that I think make brilliant first dogs.<br><br>Every trainer and dog lover will tell you something different about what breeds are best for you. At the end of the day, it’s your choice. But these are my top five dogs for a first-time owner. I’ve chosen them based on a decade’s experience of working with breeds of all sorts and seeing firsthand some of the common problems among dogs. These five are all typically easygoing, good-natured, smart, and willing to learn. The Rottweiler man in me can observe occasional “over-friendliness” in these breeds, but that’s not a bad thing for beginners, and basically makes them perfect for the novice trainer. If your heart is set on an American bully, but you’ve never had a dog before, think about having one of these dogs first—you can always grow your family l...</code> |
* Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters:
```json
{
"loss": "MultipleNegativesRankingLoss",
"matryoshka_dims": [
1024,
512,
256,
128,
64
],
"matryoshka_weights": [
1,
1,
1,
1,
1
],
"n_dims_per_step": -1
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `eval_strategy`: steps
- `per_device_train_batch_size`: 16
- `per_device_eval_batch_size`: 16
- `num_train_epochs`: 5
- `multi_dataset_batch_sampler`: round_robin
#### All Hyperparameters
<details><summary>Click to expand</summary>
- `overwrite_output_dir`: False
- `do_predict`: False
- `eval_strategy`: steps
- `prediction_loss_only`: True
- `per_device_train_batch_size`: 16
- `per_device_eval_batch_size`: 16
- `per_gpu_train_batch_size`: None
- `per_gpu_eval_batch_size`: None
- `gradient_accumulation_steps`: 1
- `eval_accumulation_steps`: None
- `torch_empty_cache_steps`: None
- `learning_rate`: 5e-05
- `weight_decay`: 0.0
- `adam_beta1`: 0.9
- `adam_beta2`: 0.999
- `adam_epsilon`: 1e-08
- `max_grad_norm`: 1
- `num_train_epochs`: 5
- `max_steps`: -1
- `lr_scheduler_type`: linear
- `lr_scheduler_kwargs`: {}
- `warmup_ratio`: 0.0
- `warmup_steps`: 0
- `log_level`: passive
- `log_level_replica`: warning
- `log_on_each_node`: True
- `logging_nan_inf_filter`: True
- `save_safetensors`: True
- `save_on_each_node`: False
- `save_only_model`: False
- `restore_callback_states_from_checkpoint`: False
- `no_cuda`: False
- `use_cpu`: False
- `use_mps_device`: False
- `seed`: 42
- `data_seed`: None
- `jit_mode_eval`: False
- `use_ipex`: False
- `bf16`: False
- `fp16`: False
- `fp16_opt_level`: O1
- `half_precision_backend`: auto
- `bf16_full_eval`: False
- `fp16_full_eval`: False
- `tf32`: None
- `local_rank`: 0
- `ddp_backend`: None
- `tpu_num_cores`: None
- `tpu_metrics_debug`: False
- `debug`: []
- `dataloader_drop_last`: False
- `dataloader_num_workers`: 0
- `dataloader_prefetch_factor`: None
- `past_index`: -1
- `disable_tqdm`: False
- `remove_unused_columns`: True
- `label_names`: None
- `load_best_model_at_end`: False
- `ignore_data_skip`: False
- `fsdp`: []
- `fsdp_min_num_params`: 0
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
- `fsdp_transformer_layer_cls_to_wrap`: None
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
- `deepspeed`: None
- `label_smoothing_factor`: 0.0
- `optim`: adamw_torch
- `optim_args`: None
- `adafactor`: False
- `group_by_length`: False
- `length_column_name`: length
- `ddp_find_unused_parameters`: None
- `ddp_bucket_cap_mb`: None
- `ddp_broadcast_buffers`: False
- `dataloader_pin_memory`: True
- `dataloader_persistent_workers`: False
- `skip_memory_metrics`: True
- `use_legacy_prediction_loop`: False
- `push_to_hub`: False
- `resume_from_checkpoint`: None
- `hub_model_id`: None
- `hub_strategy`: every_save
- `hub_private_repo`: None
- `hub_always_push`: False
- `gradient_checkpointing`: False
- `gradient_checkpointing_kwargs`: None
- `include_inputs_for_metrics`: False
- `include_for_metrics`: []
- `eval_do_concat_batches`: True
- `fp16_backend`: auto
- `push_to_hub_model_id`: None
- `push_to_hub_organization`: None
- `mp_parameters`:
- `auto_find_batch_size`: False
- `full_determinism`: False
- `torchdynamo`: None
- `ray_scope`: last
- `ddp_timeout`: 1800
- `torch_compile`: False
- `torch_compile_backend`: None
- `torch_compile_mode`: None
- `include_tokens_per_second`: False
- `include_num_input_tokens_seen`: False
- `neftune_noise_alpha`: None
- `optim_target_modules`: None
- `batch_eval_metrics`: False
- `eval_on_start`: False
- `use_liger_kernel`: False
- `eval_use_gather_object`: False
- `average_tokens_across_devices`: False
- `prompts`: None
- `batch_sampler`: batch_sampler
- `multi_dataset_batch_sampler`: round_robin
</details>
### Training Logs
| Epoch | Step | cosine_ndcg@10 |
|:------:|:----:|:--------------:|
| 0.5682 | 25 | 0.7986 |
| 1.0 | 44 | 0.8182 |
| 1.1364 | 50 | 0.8224 |
| 1.7045 | 75 | 0.8181 |
| 2.0 | 88 | 0.8224 |
| 2.2727 | 100 | 0.8205 |
| 2.8409 | 125 | 0.8221 |
| 3.0 | 132 | 0.8235 |
| 3.4091 | 150 | 0.8205 |
| 3.9773 | 175 | 0.8178 |
| 4.0 | 176 | 0.8184 |
| 4.5455 | 200 | 0.8204 |
| 5.0 | 220 | 0.8202 |
### Framework Versions
- Python: 3.11.12
- Sentence Transformers: 4.1.0
- Transformers: 4.52.2
- PyTorch: 2.6.0+cu124
- Accelerate: 1.7.0
- Datasets: 2.14.4
- Tokenizers: 0.21.1
## Citation
### BibTeX
#### Sentence Transformers
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
```
#### MatryoshkaLoss
```bibtex
@misc{kusupati2024matryoshka,
title={Matryoshka Representation Learning},
author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
year={2024},
eprint={2205.13147},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
```
#### MultipleNegativesRankingLoss
```bibtex
@misc{henderson2017efficient,
title={Efficient Natural Language Response Suggestion for Smart Reply},
author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
year={2017},
eprint={1705.00652},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> |
Ftmhd/bart-base-finetuned-steel-news-Environment | Ftmhd | 2025-05-29T12:26:29Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"bart",
"text2text-generation",
"generated_from_trainer",
"base_model:facebook/bart-base",
"base_model:finetune:facebook/bart-base",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | text2text-generation | 2025-05-29T12:25:45Z | ---
library_name: transformers
license: apache-2.0
base_model: facebook/bart-base
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: bart-base-finetuned-steel-news-Environment
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bart-base-finetuned-steel-news-Environment
This model is a fine-tuned version of [facebook/bart-base](https://huggingface.co/facebook/bart-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 5.0878
- Rouge1: 33.7492
- Rouge2: 16.9181
- Rougel: 31.1710
- Rougelsum: 31.1290
- Rouge3: 5.2115
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 16
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 5
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Rouge3 |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:------:|
| No log | 1.0 | 23 | 13.6118 | 23.8890 | 8.6893 | 19.6100 | 19.7700 | 2.9881 |
| No log | 2.0 | 46 | 11.2290 | 23.8890 | 8.6893 | 19.6100 | 19.7700 | 2.9881 |
| No log | 3.0 | 69 | 10.0166 | 23.7559 | 8.2513 | 19.3871 | 19.4579 | 2.4654 |
| No log | 4.0 | 92 | 8.0345 | 27.0003 | 9.6629 | 24.1700 | 24.1187 | 2.0520 |
| 11.4324 | 5.0 | 115 | 4.9696 | 32.3997 | 12.4816 | 29.5265 | 29.4606 | 3.2122 |
### Framework versions
- Transformers 4.52.2
- Pytorch 2.6.0+cu124
- Datasets 2.14.4
- Tokenizers 0.21.1
|
luckysantoso/adapter-qwen3-capd-lora | luckysantoso | 2025-05-29T12:23:42Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"unsloth",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-05-29T12:19:31Z | ---
library_name: transformers
tags:
- unsloth
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
TareksTesting/Scripturient-V2.3-LLaMa-70B | TareksTesting | 2025-05-29T12:19:41Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"llama",
"text-generation",
"mergekit",
"merge",
"conversational",
"arxiv:2406.11617",
"base_model:TareksLab/Amethyst-SCE-V4-LLaMa-70B",
"base_model:merge:TareksLab/Amethyst-SCE-V4-LLaMa-70B",
"base_model:TareksLab/Carnelian-SCE-V4-LLaMa-70B",
"base_model:merge:TareksLab/Carnelian-SCE-V4-LLaMa-70B",
"base_model:TareksLab/Citrine-MS-V3-LLaMa-70B",
"base_model:merge:TareksLab/Citrine-MS-V3-LLaMa-70B",
"base_model:TareksLab/Diamond-DL-V1-LLaMa-70B",
"base_model:merge:TareksLab/Diamond-DL-V1-LLaMa-70B",
"base_model:TareksLab/Emerald-SCE-V3-LLaMa-70B",
"base_model:merge:TareksLab/Emerald-SCE-V3-LLaMa-70B",
"base_model:TareksLab/Ruby-D-V3-LLaMa-70B",
"base_model:merge:TareksLab/Ruby-D-V3-LLaMa-70B",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-29T11:44:10Z | ---
base_model:
- TareksLab/Ruby-D-V3-LLaMa-70B
- TareksLab/Citrine-MS-V3-LLaMa-70B
- TareksLab/Diamond-DL-V1-LLaMa-70B
- TareksLab/Amethyst-SCE-V4-LLaMa-70B
- TareksLab/Carnelian-SCE-V4-LLaMa-70B
- TareksLab/Emerald-SCE-V3-LLaMa-70B
library_name: transformers
tags:
- mergekit
- merge
---
# MERGE5
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
## Merge Details
### Merge Method
This model was merged using the [Linear DELLA](https://arxiv.org/abs/2406.11617) merge method using [TareksLab/Diamond-DL-V1-LLaMa-70B](https://huggingface.co/TareksLab/Diamond-DL-V1-LLaMa-70B) as a base.
### Models Merged
The following models were included in the merge:
* [TareksLab/Ruby-D-V3-LLaMa-70B](https://huggingface.co/TareksLab/Ruby-D-V3-LLaMa-70B)
* [TareksLab/Citrine-MS-V3-LLaMa-70B](https://huggingface.co/TareksLab/Citrine-MS-V3-LLaMa-70B)
* [TareksLab/Amethyst-SCE-V4-LLaMa-70B](https://huggingface.co/TareksLab/Amethyst-SCE-V4-LLaMa-70B)
* [TareksLab/Carnelian-SCE-V4-LLaMa-70B](https://huggingface.co/TareksLab/Carnelian-SCE-V4-LLaMa-70B)
* [TareksLab/Emerald-SCE-V3-LLaMa-70B](https://huggingface.co/TareksLab/Emerald-SCE-V3-LLaMa-70B)
### Configuration
The following YAML configuration was used to produce this model:
```yaml
models:
- model: TareksLab/Diamond-DL-V1-LLaMa-70B
parameters:
weight: 0.10
density: 0.7
epsilon: 0.20
- model: TareksLab/Citrine-MS-V3-LLaMa-70B
parameters:
weight: [0.35, 0.25, 0.17, 0.12, 0.11]
density: 0.7
epsilon: 0.20
- model: TareksLab/Amethyst-SCE-V4-LLaMa-70B
parameters:
weight: [0.24, 0.27, 0.2, 0.16, 0.13]
density: 0.7
epsilon: 0.20
- model: TareksLab/Ruby-D-V3-LLaMa-70B
parameters:
weight: [0.17, 0.2, 0.26, 0.2, 0.17]
density: 0.7
epsilon: 0.20
- model: TareksLab/Carnelian-SCE-V4-LLaMa-70B
parameters:
weight: [0.13, 0.16, 0.2, 0.27, 0.24]
density: 0.7
epsilon: 0.20
- model: TareksLab/Emerald-SCE-V3-LLaMa-70B
parameters:
weight: [0.11, 0.12, 0.17, 0.25, 0.35]
density: 0.7
epsilon: 0.20
merge_method: della_linear
base_model: TareksLab/Diamond-DL-V1-LLaMa-70B
parameters:
lambda: 1.1
normalize: false
dtype: float32
out_dtype: bfloat16
chat_template: llama3
tokenizer:
source: TareksLab/Ruby-D-V3-LLaMa-70B
pad_to_multiple_of: 8
```
|
Ekata/mcqa-dpo-v2 | Ekata | 2025-05-29T12:04:35Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"generated_from_trainer",
"trl",
"dpo",
"arxiv:2305.18290",
"base_model:meta-llama/Llama-3.1-8B-Instruct",
"base_model:finetune:meta-llama/Llama-3.1-8B-Instruct",
"endpoints_compatible",
"region:us"
] | null | 2025-05-29T12:02:00Z | ---
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
library_name: transformers
model_name: mcqa-dpo-v2
tags:
- generated_from_trainer
- trl
- dpo
licence: license
---
# Model Card for mcqa-dpo-v2
This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct).
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="Ekata/mcqa-dpo-v2", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
This model was trained with DPO, a method introduced in [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://huggingface.co/papers/2305.18290).
### Framework versions
- TRL: 0.17.0
- Transformers: 4.52.3
- Pytorch: 2.5.1+cu121
- Datasets: 3.2.0
- Tokenizers: 0.21.0
## Citations
Cite DPO as:
```bibtex
@inproceedings{rafailov2023direct,
title = {{Direct Preference Optimization: Your Language Model is Secretly a Reward Model}},
author = {Rafael Rafailov and Archit Sharma and Eric Mitchell and Christopher D. Manning and Stefano Ermon and Chelsea Finn},
year = 2023,
booktitle = {Advances in Neural Information Processing Systems 36: Annual Conference on Neural Information Processing Systems 2023, NeurIPS 2023, New Orleans, LA, USA, December 10 - 16, 2023},
url = {http://papers.nips.cc/paper_files/paper/2023/hash/a85b405ed65c6477a4fe8302b5e06ce7-Abstract-Conference.html},
editor = {Alice Oh and Tristan Naumann and Amir Globerson and Kate Saenko and Moritz Hardt and Sergey Levine},
}
```
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
``` |
volfenstein/wolfgang-story-generator | volfenstein | 2025-05-29T12:03:31Z | 33 | 0 | transformers | [
"transformers",
"safetensors",
"llama",
"text-generation",
"generated_from_trainer",
"trl",
"sft",
"conversational",
"base_model:SimpleStories/SimpleStories-35M",
"base_model:finetune:SimpleStories/SimpleStories-35M",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-25T15:30:13Z | ---
base_model: SimpleStories/SimpleStories-35M
library_name: transformers
model_name: wolfgang-story-generator
tags:
- generated_from_trainer
- trl
- sft
licence: license
---
# Model Card for wolfgang-story-generator
This model is a fine-tuned version of [SimpleStories/SimpleStories-35M](https://huggingface.co/SimpleStories/SimpleStories-35M).
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="volfenstein/wolfgang-story-generator", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/volfenstein-ruhr-universit-t-bochum/huggingface/runs/bhk25476)
This model was trained with SFT.
### Framework versions
- TRL: 0.18.0
- Transformers: 4.52.3
- Pytorch: 2.6.0+cu124
- Datasets: 3.6.0
- Tokenizers: 0.21.1
## Citations
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
``` |
BKM1804/Qwen2-0.5B-Instruct-7e4bf26b-d4ca-414d-b37b-1a1919ea88ef-dpo-tuned-merged-check | BKM1804 | 2025-05-29T11:52:20Z | 0 | 0 | transformers | [
"transformers",
"text-generation-inference",
"unsloth",
"qwen2",
"en",
"base_model:unsloth/Qwen2-0.5B-Instruct",
"base_model:finetune:unsloth/Qwen2-0.5B-Instruct",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-05-29T11:52:19Z | ---
base_model: unsloth/Qwen2-0.5B-Instruct
tags:
- text-generation-inference
- transformers
- unsloth
- qwen2
license: apache-2.0
language:
- en
---
# Uploaded finetuned model
- **Developed by:** BKM1804
- **License:** apache-2.0
- **Finetuned from model :** unsloth/Qwen2-0.5B-Instruct
This qwen2 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
kenohammer/lubna.qureshi.viral.video | kenohammer | 2025-05-29T11:48:11Z | 0 | 0 | null | [
"region:us"
] | null | 2025-05-29T11:47:22Z | <a href="https://lojinx.cfd/ydtsrr"> 🌐 Click Here To link (lubna.qureshi.viral.video)
🔴 ➤►DOWNLOAD👉👉🟢 ➤ <a href="https://lojinx.cfd/ydtsrr"> 🌐 lubna.qureshi.viral.video
|
mnm3/gemma-3-27b-it-mnm3-lora-adapters | mnm3 | 2025-05-29T11:45:48Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-05-29T11:45:19Z | ---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
Shubh0/kylie-flux | Shubh0 | 2025-05-29T11:28:53Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T11:12:38Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: KYLIE
---
# Kylie Flux
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `KYLIE` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "KYLIE",
"lora_weights": "https://huggingface.co/Shubh0/kylie-flux/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('Shubh0/kylie-flux', weight_name='lora.safetensors')
image = pipeline('KYLIE').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 1500
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/Shubh0/kylie-flux/discussions) to add images that show off what you’ve made with this LoRA.
|
loretyan/vit-base-oxford_flowers102 | loretyan | 2025-05-29T11:27:53Z | 0 | 0 | transformers | [
"transformers",
"tensorboard",
"safetensors",
"vit",
"image-classification",
"generated_from_trainer",
"base_model:google/vit-base-patch16-224",
"base_model:finetune:google/vit-base-patch16-224",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | image-classification | 2025-05-28T13:52:36Z | ---
library_name: transformers
license: apache-2.0
base_model: google/vit-base-patch16-224
tags:
- image-classification
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: vit-base-oxford_flowers102
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# vit-base-oxford_flowers102
This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the dpdl-benchmark/oxford_flowers102 dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1771
- Accuracy: 0.9706
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log | 1.0 | 51 | 3.4099 | 0.5490 |
| 3.341 | 2.0 | 102 | 2.3499 | 0.8431 |
| 3.341 | 3.0 | 153 | 1.6991 | 0.9020 |
| 1.3666 | 4.0 | 204 | 1.3820 | 0.9412 |
| 1.3666 | 5.0 | 255 | 1.2861 | 0.9510 |
### Framework versions
- Transformers 4.52.3
- Pytorch 2.6.0+cpu
- Datasets 3.6.0
- Tokenizers 0.21.1
### Results
Accuracy: 0.0137
Precision: 0.0017
Recall: 0.0137
|
gehatzip/Llama-3.2-1B-Q4_K_M-GGUF | gehatzip | 2025-05-29T11:27:25Z | 0 | 0 | transformers | [
"transformers",
"gguf",
"facebook",
"meta",
"pytorch",
"llama",
"llama-3",
"llama-cpp",
"gguf-my-repo",
"text-generation",
"en",
"de",
"fr",
"it",
"pt",
"hi",
"es",
"th",
"base_model:meta-llama/Llama-3.2-1B",
"base_model:quantized:meta-llama/Llama-3.2-1B",
"license:llama3.2",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-29T11:27:18Z | ---
language:
- en
- de
- fr
- it
- pt
- hi
- es
- th
library_name: transformers
pipeline_tag: text-generation
tags:
- facebook
- meta
- pytorch
- llama
- llama-3
- llama-cpp
- gguf-my-repo
license: llama3.2
extra_gated_prompt: "### LLAMA 3.2 COMMUNITY LICENSE AGREEMENT\n\nLlama 3.2 Version\
\ Release Date: September 25, 2024\n\n“Agreement” means the terms and conditions\
\ for use, reproduction, distribution and modification of the Llama Materials set\
\ forth herein.\n\n“Documentation” means the specifications, manuals and documentation\
\ accompanying Llama 3.2 distributed by Meta at https://llama.meta.com/doc/overview.\n\
\n“Licensee” or “you” means you, or your employer or any other person or entity\
\ (if you are entering into this Agreement on such person or entity’s behalf),\
\ of the age required under applicable laws, rules or regulations to provide legal\
\ consent and that has legal authority to bind your employer or such other person\
\ or entity if you are entering in this Agreement on their behalf.\n\n“Llama 3.2”\
\ means the foundational large language models and software and algorithms, including\
\ machine-learning model code, trained model weights, inference-enabling code, training-enabling\
\ code, fine-tuning enabling code and other elements of the foregoing distributed\
\ by Meta at https://www.llama.com/llama-downloads.\n\n“Llama Materials” means,\
\ collectively, Meta’s proprietary Llama 3.2 and Documentation (and any portion\
\ thereof) made available under this Agreement.\n\n“Meta” or “we” means Meta Platforms\
\ Ireland Limited (if you are located in or, if you are an entity, your principal\
\ place of business is in the EEA or Switzerland) and Meta Platforms, Inc. (if\
\ you are located outside of the EEA or Switzerland). \n\nBy clicking “I Accept”\
\ below or by using or distributing any portion or element of the Llama Materials,\
\ you agree to be bound by this Agreement.\n\n1. License Rights and Redistribution.\n\
a. Grant of Rights. You are granted a non-exclusive, worldwide, non-transferable\
\ and royalty-free limited license under Meta’s intellectual property or other rights\
\ owned by Meta embodied in the Llama Materials to use, reproduce, distribute,\
\ copy, create derivative works of, and make modifications to the Llama Materials.\
\ \nb. Redistribution and Use. \ni. If you distribute or make available the Llama\
\ Materials (or any derivative works thereof), or a product or service (including\
\ another AI model) that contains any of them, you shall (A) provide a copy of this\
\ Agreement with any such Llama Materials; and (B) prominently display “Built with\
\ Llama” on a related website, user interface, blogpost, about page, or product\
\ documentation. If you use the Llama Materials or any outputs or results of the\
\ Llama Materials to create, train, fine tune, or otherwise improve an AI model,\
\ which is distributed or made available, you shall also include “Llama” at the\
\ beginning of any such AI model name.\nii. If you receive Llama Materials, or any\
\ derivative works thereof, from a Licensee as part of an integrated end user product,\
\ then Section 2 of this Agreement will not apply to you. \niii. You must retain\
\ in all copies of the Llama Materials that you distribute the following attribution\
\ notice within a “Notice” text file distributed as a part of such copies: “Llama\
\ 3.2 is licensed under the Llama 3.2 Community License, Copyright © Meta Platforms,\
\ Inc. All Rights Reserved.”\niv. Your use of the Llama Materials must comply with\
\ applicable laws and regulations (including trade compliance laws and regulations)\
\ and adhere to the Acceptable Use Policy for the Llama Materials (available at\
\ https://www.llama.com/llama3_2/use-policy), which is hereby incorporated by reference\
\ into this Agreement.\n \n2. Additional Commercial Terms. If, on the Llama 3.2\
\ version release date, the monthly active users of the products or services made\
\ available by or for Licensee, or Licensee’s affiliates, is greater than 700 million\
\ monthly active users in the preceding calendar month, you must request a license\
\ from Meta, which Meta may grant to you in its sole discretion, and you are not\
\ authorized to exercise any of the rights under this Agreement unless or until\
\ Meta otherwise expressly grants you such rights.\n3. Disclaimer of Warranty. UNLESS\
\ REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS THEREFROM\
\ ARE PROVIDED ON AN “AS IS” BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS\
\ ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION,\
\ ANY WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR\
\ PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING\
\ OR REDISTRIBUTING THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR\
\ USE OF THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS.\n4. Limitation of Liability.\
\ IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY,\
\ WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING\
\ OUT OF THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL,\
\ INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE\
\ BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.\n5. Intellectual Property.\n\
a. No trademark licenses are granted under this Agreement, and in connection with\
\ the Llama Materials, neither Meta nor Licensee may use any name or mark owned\
\ by or associated with the other or any of its affiliates, except as required\
\ for reasonable and customary use in describing and redistributing the Llama Materials\
\ or as set forth in this Section 5(a). Meta hereby grants you a license to use\
\ “Llama” (the “Mark”) solely as required to comply with the last sentence of Section\
\ 1.b.i. You will comply with Meta’s brand guidelines (currently accessible at\
\ https://about.meta.com/brand/resources/meta/company-brand/). All goodwill arising\
\ out of your use of the Mark will inure to the benefit of Meta.\nb. Subject to\
\ Meta’s ownership of Llama Materials and derivatives made by or for Meta, with\
\ respect to any derivative works and modifications of the Llama Materials that\
\ are made by you, as between you and Meta, you are and will be the owner of such\
\ derivative works and modifications.\nc. If you institute litigation or other proceedings\
\ against Meta or any entity (including a cross-claim or counterclaim in a lawsuit)\
\ alleging that the Llama Materials or Llama 3.2 outputs or results, or any portion\
\ of any of the foregoing, constitutes infringement of intellectual property or\
\ other rights owned or licensable by you, then any licenses granted to you under\
\ this Agreement shall terminate as of the date such litigation or claim is filed\
\ or instituted. You will indemnify and hold harmless Meta from and against any\
\ claim by any third party arising out of or related to your use or distribution\
\ of the Llama Materials.\n6. Term and Termination. The term of this Agreement will\
\ commence upon your acceptance of this Agreement or access to the Llama Materials\
\ and will continue in full force and effect until terminated in accordance with\
\ the terms and conditions herein. Meta may terminate this Agreement if you are\
\ in breach of any term or condition of this Agreement. Upon termination of this\
\ Agreement, you shall delete and cease use of the Llama Materials. Sections 3,\
\ 4 and 7 shall survive the termination of this Agreement. \n7. Governing Law and\
\ Jurisdiction. This Agreement will be governed and construed under the laws of\
\ the State of California without regard to choice of law principles, and the UN\
\ Convention on Contracts for the International Sale of Goods does not apply to\
\ this Agreement. The courts of California shall have exclusive jurisdiction of\
\ any dispute arising out of this Agreement. \n### Llama 3.2 Acceptable Use Policy\n\
Meta is committed to promoting safe and fair use of its tools and features, including\
\ Llama 3.2. If you access or use Llama 3.2, you agree to this Acceptable Use Policy\
\ (“**Policy**”). The most recent copy of this policy can be found at [https://www.llama.com/llama3_2/use-policy](https://www.llama.com/llama3_2/use-policy).\n\
#### Prohibited Uses\nWe want everyone to use Llama 3.2 safely and responsibly.\
\ You agree you will not use, or allow others to use, Llama 3.2 to:\n1. Violate\
\ the law or others’ rights, including to:\n 1. Engage in, promote, generate,\
\ contribute to, encourage, plan, incite, or further illegal or unlawful activity\
\ or content, such as:\n 1. Violence or terrorism\n 2. Exploitation\
\ or harm to children, including the solicitation, creation, acquisition, or dissemination\
\ of child exploitative content or failure to report Child Sexual Abuse Material\n\
\ 3. Human trafficking, exploitation, and sexual violence\n 4. The\
\ illegal distribution of information or materials to minors, including obscene\
\ materials, or failure to employ legally required age-gating in connection with\
\ such information or materials.\n 5. Sexual solicitation\n 6. Any\
\ other criminal activity\n 1. Engage in, promote, incite, or facilitate the\
\ harassment, abuse, threatening, or bullying of individuals or groups of individuals\n\
\ 2. Engage in, promote, incite, or facilitate discrimination or other unlawful\
\ or harmful conduct in the provision of employment, employment benefits, credit,\
\ housing, other economic benefits, or other essential goods and services\n 3.\
\ Engage in the unauthorized or unlicensed practice of any profession including,\
\ but not limited to, financial, legal, medical/health, or related professional\
\ practices\n 4. Collect, process, disclose, generate, or infer private or sensitive\
\ information about individuals, including information about individuals’ identity,\
\ health, or demographic information, unless you have obtained the right to do so\
\ in accordance with applicable law\n 5. Engage in or facilitate any action or\
\ generate any content that infringes, misappropriates, or otherwise violates any\
\ third-party rights, including the outputs or results of any products or services\
\ using the Llama Materials\n 6. Create, generate, or facilitate the creation\
\ of malicious code, malware, computer viruses or do anything else that could disable,\
\ overburden, interfere with or impair the proper working, integrity, operation\
\ or appearance of a website or computer system\n 7. Engage in any action, or\
\ facilitate any action, to intentionally circumvent or remove usage restrictions\
\ or other safety measures, or to enable functionality disabled by Meta \n2. Engage\
\ in, promote, incite, facilitate, or assist in the planning or development of activities\
\ that present a risk of death or bodily harm to individuals, including use of Llama\
\ 3.2 related to the following:\n 8. Military, warfare, nuclear industries or\
\ applications, espionage, use for materials or activities that are subject to the\
\ International Traffic Arms Regulations (ITAR) maintained by the United States\
\ Department of State or to the U.S. Biological Weapons Anti-Terrorism Act of 1989\
\ or the Chemical Weapons Convention Implementation Act of 1997\n 9. Guns and\
\ illegal weapons (including weapon development)\n 10. Illegal drugs and regulated/controlled\
\ substances\n 11. Operation of critical infrastructure, transportation technologies,\
\ or heavy machinery\n 12. Self-harm or harm to others, including suicide, cutting,\
\ and eating disorders\n 13. Any content intended to incite or promote violence,\
\ abuse, or any infliction of bodily harm to an individual\n3. Intentionally deceive\
\ or mislead others, including use of Llama 3.2 related to the following:\n 14.\
\ Generating, promoting, or furthering fraud or the creation or promotion of disinformation\n\
\ 15. Generating, promoting, or furthering defamatory content, including the\
\ creation of defamatory statements, images, or other content\n 16. Generating,\
\ promoting, or further distributing spam\n 17. Impersonating another individual\
\ without consent, authorization, or legal right\n 18. Representing that the\
\ use of Llama 3.2 or outputs are human-generated\n 19. Generating or facilitating\
\ false online engagement, including fake reviews and other means of fake online\
\ engagement \n4. Fail to appropriately disclose to end users any known dangers\
\ of your AI system 5. Interact with third party tools, models, or software designed\
\ to generate unlawful content or engage in unlawful or harmful conduct and/or represent\
\ that the outputs of such tools, models, or software are associated with Meta or\
\ Llama 3.2\n\nWith respect to any multimodal models included in Llama 3.2, the\
\ rights granted under Section 1(a) of the Llama 3.2 Community License Agreement\
\ are not being granted to you if you are an individual domiciled in, or a company\
\ with a principal place of business in, the European Union. This restriction does\
\ not apply to end users of a product or service that incorporates any such multimodal\
\ models.\n\nPlease report any violation of this Policy, software “bug,” or other\
\ problems that could lead to a violation of this Policy through one of the following\
\ means:\n\n* Reporting issues with the model: [https://github.com/meta-llama/llama-models/issues](https://l.workplace.com/l.php?u=https%3A%2F%2Fgithub.com%2Fmeta-llama%2Fllama-models%2Fissues&h=AT0qV8W9BFT6NwihiOHRuKYQM_UnkzN_NmHMy91OT55gkLpgi4kQupHUl0ssR4dQsIQ8n3tfd0vtkobvsEvt1l4Ic6GXI2EeuHV8N08OG2WnbAmm0FL4ObkazC6G_256vN0lN9DsykCvCqGZ)\n\
* Reporting risky content generated by the model: [developers.facebook.com/llama_output_feedback](http://developers.facebook.com/llama_output_feedback)\n\
* Reporting bugs and security concerns: [facebook.com/whitehat/info](http://facebook.com/whitehat/info)\n\
* Reporting violations of the Acceptable Use Policy or unlicensed uses of Llama\
\ 3.2: [email protected]"
extra_gated_fields:
First Name: text
Last Name: text
Date of birth: date_picker
Country: country
Affiliation: text
Job title:
type: select
options:
- Student
- Research Graduate
- AI researcher
- AI developer/engineer
- Reporter
- Other
geo: ip_location
? By clicking Submit below I accept the terms of the license and acknowledge that
the information I provide will be collected stored processed and shared in accordance
with the Meta Privacy Policy
: checkbox
extra_gated_description: The information you provide will be collected, stored, processed
and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/).
extra_gated_button_content: Submit
base_model: meta-llama/Llama-3.2-1B
---
# gehatzip/Llama-3.2-1B-Q4_K_M-GGUF
This model was converted to GGUF format from [`meta-llama/Llama-3.2-1B`](https://huggingface.co/meta-llama/Llama-3.2-1B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
Refer to the [original model card](https://huggingface.co/meta-llama/Llama-3.2-1B) for more details on the model.
## Use with llama.cpp
Install llama.cpp through brew (works on Mac and Linux)
```bash
brew install llama.cpp
```
Invoke the llama.cpp server or the CLI.
### CLI:
```bash
llama-cli --hf-repo gehatzip/Llama-3.2-1B-Q4_K_M-GGUF --hf-file llama-3.2-1b-q4_k_m.gguf -p "The meaning to life and the universe is"
```
### Server:
```bash
llama-server --hf-repo gehatzip/Llama-3.2-1B-Q4_K_M-GGUF --hf-file llama-3.2-1b-q4_k_m.gguf -c 2048
```
Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well.
Step 1: Clone llama.cpp from GitHub.
```
git clone https://github.com/ggerganov/llama.cpp
```
Step 2: Move into the llama.cpp folder and build it with `LLAMA_CURL=1` flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux).
```
cd llama.cpp && LLAMA_CURL=1 make
```
Step 3: Run inference through the main binary.
```
./llama-cli --hf-repo gehatzip/Llama-3.2-1B-Q4_K_M-GGUF --hf-file llama-3.2-1b-q4_k_m.gguf -p "The meaning to life and the universe is"
```
or
```
./llama-server --hf-repo gehatzip/Llama-3.2-1B-Q4_K_M-GGUF --hf-file llama-3.2-1b-q4_k_m.gguf -c 2048
```
|
BeckerAnas/wise-cosmos-218 | BeckerAnas | 2025-05-29T11:27:03Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"convnextv2",
"image-classification",
"generated_from_trainer",
"base_model:facebook/convnextv2-tiny-1k-224",
"base_model:finetune:facebook/convnextv2-tiny-1k-224",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | image-classification | 2025-05-29T08:31:36Z | ---
library_name: transformers
license: apache-2.0
base_model: facebook/convnextv2-tiny-1k-224
tags:
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: wise-cosmos-218
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# wise-cosmos-218
This model is a fine-tuned version of [facebook/convnextv2-tiny-1k-224](https://huggingface.co/facebook/convnextv2-tiny-1k-224) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5895
- Accuracy: 0.6328
- Precision: 0.6756
- Recall: 0.6328
- F1: 0.6413
- Roc Auc: 0.8748
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 256
- eval_batch_size: 256
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | Roc Auc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|:-------:|
| 1.3576 | 1.0 | 17 | 1.2560 | 0.2383 | 0.4653 | 0.2383 | 0.3033 | 0.7652 |
| 1.0327 | 2.0 | 34 | 0.8853 | 0.5326 | 0.5906 | 0.5326 | 0.5534 | 0.8194 |
| 0.764 | 3.0 | 51 | 0.6813 | 0.6159 | 0.6415 | 0.6159 | 0.6239 | 0.8545 |
| 0.6357 | 4.0 | 68 | 0.6066 | 0.6367 | 0.6764 | 0.6367 | 0.6464 | 0.8702 |
| 0.5797 | 5.0 | 85 | 0.5895 | 0.6328 | 0.6756 | 0.6328 | 0.6413 | 0.8748 |
### Framework versions
- Transformers 4.52.3
- Pytorch 2.7.0+cpu
- Datasets 3.6.0
- Tokenizers 0.21.0
|
OpenGVLab/InternVL3-38B-AWQ | OpenGVLab | 2025-05-29T11:25:33Z | 8,875 | 2 | transformers | [
"transformers",
"pytorch",
"internvl_chat",
"feature-extraction",
"internvl",
"custom_code",
"image-text-to-text",
"conversational",
"multilingual",
"dataset:OpenGVLab/MMPR-v1.2",
"arxiv:2312.14238",
"arxiv:2404.16821",
"arxiv:2412.05271",
"arxiv:2411.10442",
"arxiv:2504.10479",
"arxiv:2412.09616",
"base_model:OpenGVLab/InternVL3-38B",
"base_model:quantized:OpenGVLab/InternVL3-38B",
"license:other",
"region:us"
] | image-text-to-text | 2025-04-17T09:53:59Z | ---
license: other
license_name: qwen
license_link: https://huggingface.co/Qwen/Qwen2.5-72B-Instruct/blob/main/LICENSE
pipeline_tag: image-text-to-text
library_name: transformers
base_model:
- OpenGVLab/InternVL3-38B
base_model_relation: quantized
datasets:
- OpenGVLab/MMPR-v1.2
language:
- multilingual
tags:
- internvl
- custom_code
---
# InternVL3-38B
[\[📂 GitHub\]](https://github.com/OpenGVLab/InternVL) [\[📜 InternVL 1.0\]](https://huggingface.co/papers/2312.14238) [\[📜 InternVL 1.5\]](https://huggingface.co/papers/2404.16821) [\[📜 InternVL 2.5\]](https://huggingface.co/papers/2412.05271) [\[📜 InternVL2.5-MPO\]](https://huggingface.co/papers/2411.10442) [\[📜 InternVL3\]](https://huggingface.co/papers/2504.10479)
[\[🆕 Blog\]](https://internvl.github.io/blog/) [\[🗨️ Chat Demo\]](https://internvl.opengvlab.com/) [\[🤗 HF Demo\]](https://huggingface.co/spaces/OpenGVLab/InternVL) [\[🚀 Quick Start\]](#quick-start) [\[📖 Documents\]](https://internvl.readthedocs.io/en/latest/)
<div align="center">
<img width="500" alt="image" src="https://cdn-uploads.huggingface.co/production/uploads/64006c09330a45b03605bba3/zJsd2hqd3EevgXo6fNgC-.png">
</div>
## Introduction
We introduce InternVL3, an advanced multimodal large language model (MLLM) series that demonstrates superior overall performance.
Compared to InternVL 2.5, InternVL3 exhibits superior multimodal perception and reasoning capabilities, while further extending its multimodal capabilities to encompass tool usage, GUI agents, industrial image analysis, 3D vision perception, and more.
Additionally, we compare InternVL3 with Qwen2.5 Chat models, whose corresponding pre-trained base models are employed as the initialization of the langauge component in InternVL3. Benefitting from Native Multimodal Pre-Training, the InternVL3 series achieves even better overall text performance than the Qwen2.5 series.
## InternVL3 Family
In the following table, we provide an overview of the InternVL3 series.
| Model Name | Vision Part | Language Part | HF Link |
| :-----------: | :-------------------------------------------------------------------------------------: | :----------------------------------------------------------------------------: | :------------------------------------------------------: |
| InternVL3-1B | [InternViT-300M-448px-V2_5](https://huggingface.co/OpenGVLab/InternViT-300M-448px-V2_5) | [Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B) | [🤗 link](https://huggingface.co/OpenGVLab/InternVL3-1B) |
| InternVL3-2B | [InternViT-300M-448px-V2_5](https://huggingface.co/OpenGVLab/InternViT-300M-448px-V2_5) | [Qwen2.5-1.5B](https://huggingface.co/Qwen/Qwen2.5-1.5B) | [🤗 link](https://huggingface.co/OpenGVLab/InternVL3-2B) |
| InternVL3-8B | [InternViT-300M-448px-V2_5](https://huggingface.co/OpenGVLab/InternViT-300M-448px-V2_5) | [Qwen2.5-7B](https://huggingface.co/Qwen/Qwen2.5-7B) | [🤗 link](https://huggingface.co/OpenGVLab/InternVL3-8B) |
| InternVL3-9B | [InternViT-300M-448px-V2_5](https://huggingface.co/OpenGVLab/InternViT-300M-448px-V2_5) | [internlm3-8b-instruct](https://huggingface.co/internlm/internlm3-8b-instruct) | [🤗 link](https://huggingface.co/OpenGVLab/InternVL3-9B) |
| InternVL3-14B | [InternViT-300M-448px-V2_5](https://huggingface.co/OpenGVLab/InternViT-300M-448px-V2_5) | [Qwen2.5-14B](https://huggingface.co/Qwen/Qwen2.5-14B) | [🤗 link](https://huggingface.co/OpenGVLab/InternVL3-14B) |
| InternVL3-38B | [InternViT-6B-448px-V2_5](https://huggingface.co/OpenGVLab/InternViT-6B-448px-V2_5) | [Qwen2.5-32B](https://huggingface.co/Qwen/Qwen2.5-32B) | [🤗 link](https://huggingface.co/OpenGVLab/InternVL3-38B) |
| InternVL3-78B | [InternViT-6B-448px-V2_5](https://huggingface.co/OpenGVLab/InternViT-6B-448px-V2_5) | [Qwen2.5-72B](https://huggingface.co/Qwen/Qwen2.5-72B) | [🤗 link](https://huggingface.co/OpenGVLab/InternVL3-78B) |
## Model Architecture
As shown in the following figure, [InternVL3](https://internvl.github.io/blog/2025-04-11-InternVL-3/) retains the same model architecture as [InternVL 2.5](https://internvl.github.io/blog/2024-12-05-InternVL-2.5/) and its predecessors, InternVL 1.5 and 2.0, following the "ViT-MLP-LLM" paradigm. In this new version, we integrate a newly incrementally pre-trained InternViT with various pre-trained LLMs, including InternLM 3 and Qwen 2.5, using a randomly initialized MLP projector.
As in the previous version, we applied a pixel unshuffle operation, reducing the number of visual tokens to one-quarter of the original. Besides, we adopted a similar dynamic resolution strategy as InternVL 1.5, dividing images into tiles of 448×448 pixels. The key difference, starting from InternVL 2.0, is that we additionally introduced support for multi-image and video data.
Notably, in InternVL3, we integrate the [Variable Visual Position Encoding (V2PE)](https://arxiv.org/abs/2412.09616), which utilizes smaller, more flexible position increments for visual tokens. Benefiting from V2PE, InternVL3 exhibits better long context understanding capabilities compared to its predecessors.
## Training Strategy
### Native Multimodal Pre-Training
We propose a [Native Multimodal Pre-Training](https://huggingface.co/papers/2504.10479) approach that consolidates language and vision learning into a single pre-training stage.
In contrast to standard paradigms that first train a language-only model and subsequently adapt it to handle additional modalities, our method interleaves multimodal data (e.g., image-text, video-text, or image-text interleaved sequences) with large-scale textual corpora. This unified training scheme allows the model to learn both linguistic and multimodal representations simultaneously, ultimately enhancing its capability to handle vision-language tasks without the need for separate alignment or bridging modules.
Please see [our paper](https://huggingface.co/papers/2504.10479) for more details.
### Supervised Fine-Tuning
In this phase, the techniques of random JPEG compression, square loss re-weighting, and multimodal data packing proposed in [InternVL2.5](https://arxiv.org/abs/2412.05271) are also employed in the InternVL3 series.
The main advancement of the SFT phase in InternVL3 compared to InternVL2.5 lies in the use of higher-quality and more diverse training data.
Specifically, we further extend training samples for tool use, 3D scene understanding, GUI operations, long context tasks, video understanding, scientific diagrams, creative writing, and multimodal reasoning.
### Mixed Preference Optimization
During Pre-training and SFT, the model is trained to predict the next token conditioned on previous ground-truth tokens.
However, during inference, the model predicts each token based on its own prior outputs.
This discrepancy between ground-truth tokens and model-predicted tokens introduces a distribution shift, which can impair the model’s Chain-of-Thought (CoT) reasoning capabilities.
To mitigate this issue, we employ [MPO](https://arxiv.org/abs/2411.10442), which introduces additional supervision from both positive and negative samples to align the model response distribution with the ground-truth distribution, thereby improving reasoning performance.
Specifically, the training objective of MPO is a combination of
preference loss \\(\mathcal{L}_{\text{p}}\\),
quality loss \\(\mathcal{L}_{\text{q}}\\),
and generation loss \\(\mathcal{L}_{\text{g}}\\),
which can be formulated as follows:
$$
\mathcal{L}=w_{p}\cdot\mathcal{L}_{\text{p}} + w_{q}\cdot\mathcal{L}_{\text{q}} + w_{g}\cdot\mathcal{L}_{\text{g}},
$$
where \\(w_{*}\\) represents the weight assigned to each loss component. Please see [our paper](https://arxiv.org/abs/2411.10442) for more details about MPO.
### Test-Time Scaling
Test-Time Scaling has been shown to be an effective method to enhance the reasoning abilities of LLMs and MLLMs.
In this work, we use the Best-of-N evaluation strategy and employ [VisualPRM-8B](https://huggingface.co/OpenGVLab/VisualPRM-8B) as the critic model to select the best response for reasoning and mathematics evaluation.
## Evaluation on Multimodal Capability
### Multimodal Reasoning and Mathematics
### OCR, Chart, and Document Understanding
### Multi-Image & Real-World Comprehension
### Comprehensive Multimodal & Hallucination Evaluation
### Visual Grounding
### Multimodal Multilingual Understanding
### Video Understanding
### GUI Grounding
### Spatial Reasoning
## Evaluation on Language Capability
We compare InternVL3 with Qwen2.5 Chat models, whose corresponding pre-trained base models are employed as the initialization of the langauge component in InternVL3.
Benefitting from Native Multimodal Pre-Training, the InternVL3 series achieves even better overall text performance than the Qwen2.5 series.
Please note that the evaluation scores of Qwen2.5 series may differ from those officially reported, as we have adopted the prompt versions provided in the table across all datasets for OpenCompass evaluation.
## Ablation Study
### Native Multimodal Pre-Training
We conduct experiments on the InternVL2-8B model while keeping its architecture, initialization parameters, and training data entirely unchanged. Traditionally, InternVL2-8B employs a training pipeline that begins with an MLP warmup phase for feature alignment followed by an Instruction Tuning stage. In our experiments, we substitute the conventional MLP warmup phase with a native multimodal pre-training process. This modification isolates the contribution of native multimodal pre-training to the overall multimodal capability of the model.
The evaluation results in the Figure below shows that the model with native multimodal pre-training exhibits performance on most benchmarks that is comparable to the fully multi-stage-trained InternVL2-8B baseline. Furthermore, when followed by instruction tuning on higher-quality data, the model demonstrates further performance gains across evaluated multimodal tasks. These findings underscore the efficiency of native multimodal pre-training in imparting powerful multimodal capabilities to MLLMs.
### Mixed Preference Optimization
As shown in the table below, models fine-tuned with MPO demonstrate superior reasoning performance across seven multimodal reasoning benchmarks compared to their counterparts without MPO. Specifically, InternVL3-78B and InternVL3-38B outperform their counterparts by 4.1 and 4.5 points, respectively. Notably, the training data used for MPO is a subset of that used for SFT, indicating that the performance improvements primarily stem from the training algorithm rather than the training data.
### Variable Visual Position Encoding
As reported in the table below, the introduction of V2PE leads to significant performance gains across most evaluation metrics. In addition, our ablation studies—by varying the positional increment \\( \delta \\)—reveal that even for tasks primarily involving conventional contexts, relatively small \\( \delta \\) values can achieve optimal performance. These findings provide important insights for future efforts aimed at refining position encoding strategies for visual tokens in MLLMs.
## Quick Start
We provide an example code to run `InternVL3-38B` using `transformers`.
> Please use transformers>=4.37.2 to ensure the model works normally.
### Model Loading
#### 16-bit (bf16 / fp16)
```python
import torch
from transformers import AutoTokenizer, AutoModel
path = "OpenGVLab/InternVL3-38B"
model = AutoModel.from_pretrained(
path,
torch_dtype=torch.bfloat16,
low_cpu_mem_usage=True,
use_flash_attn=True,
trust_remote_code=True).eval().cuda()
```
#### BNB 8-bit Quantization
```python
import torch
from transformers import AutoTokenizer, AutoModel
path = "OpenGVLab/InternVL3-38B"
model = AutoModel.from_pretrained(
path,
torch_dtype=torch.bfloat16,
load_in_8bit=True,
low_cpu_mem_usage=True,
use_flash_attn=True,
trust_remote_code=True).eval()
```
#### Multiple GPUs
The reason for writing the code this way is to avoid errors that occur during multi-GPU inference due to tensors not being on the same device. By ensuring that the first and last layers of the large language model (LLM) are on the same device, we prevent such errors.
```python
import math
import torch
from transformers import AutoTokenizer, AutoModel
def split_model(model_name):
device_map = {}
world_size = torch.cuda.device_count()
config = AutoConfig.from_pretrained(model_path, trust_remote_code=True)
num_layers = config.llm_config.num_hidden_layers
# Since the first GPU will be used for ViT, treat it as half a GPU.
num_layers_per_gpu = math.ceil(num_layers / (world_size - 0.5))
num_layers_per_gpu = [num_layers_per_gpu] * world_size
num_layers_per_gpu[0] = math.ceil(num_layers_per_gpu[0] * 0.5)
layer_cnt = 0
for i, num_layer in enumerate(num_layers_per_gpu):
for j in range(num_layer):
device_map[f'language_model.model.layers.{layer_cnt}'] = i
layer_cnt += 1
device_map['vision_model'] = 0
device_map['mlp1'] = 0
device_map['language_model.model.tok_embeddings'] = 0
device_map['language_model.model.embed_tokens'] = 0
device_map['language_model.output'] = 0
device_map['language_model.model.norm'] = 0
device_map['language_model.model.rotary_emb'] = 0
device_map['language_model.lm_head'] = 0
device_map[f'language_model.model.layers.{num_layers - 1}'] = 0
return device_map
path = "OpenGVLab/InternVL3-38B"
device_map = split_model('InternVL3-38B')
model = AutoModel.from_pretrained(
path,
torch_dtype=torch.bfloat16,
low_cpu_mem_usage=True,
use_flash_attn=True,
trust_remote_code=True,
device_map=device_map).eval()
```
### Inference with Transformers
```python
import math
import numpy as np
import torch
import torchvision.transforms as T
from decord import VideoReader, cpu
from PIL import Image
from torchvision.transforms.functional import InterpolationMode
from transformers import AutoModel, AutoTokenizer
IMAGENET_MEAN = (0.485, 0.456, 0.406)
IMAGENET_STD = (0.229, 0.224, 0.225)
def build_transform(input_size):
MEAN, STD = IMAGENET_MEAN, IMAGENET_STD
transform = T.Compose([
T.Lambda(lambda img: img.convert('RGB') if img.mode != 'RGB' else img),
T.Resize((input_size, input_size), interpolation=InterpolationMode.BICUBIC),
T.ToTensor(),
T.Normalize(mean=MEAN, std=STD)
])
return transform
def find_closest_aspect_ratio(aspect_ratio, target_ratios, width, height, image_size):
best_ratio_diff = float('inf')
best_ratio = (1, 1)
area = width * height
for ratio in target_ratios:
target_aspect_ratio = ratio[0] / ratio[1]
ratio_diff = abs(aspect_ratio - target_aspect_ratio)
if ratio_diff < best_ratio_diff:
best_ratio_diff = ratio_diff
best_ratio = ratio
elif ratio_diff == best_ratio_diff:
if area > 0.5 * image_size * image_size * ratio[0] * ratio[1]:
best_ratio = ratio
return best_ratio
def dynamic_preprocess(image, min_num=1, max_num=12, image_size=448, use_thumbnail=False):
orig_width, orig_height = image.size
aspect_ratio = orig_width / orig_height
# calculate the existing image aspect ratio
target_ratios = set(
(i, j) for n in range(min_num, max_num + 1) for i in range(1, n + 1) for j in range(1, n + 1) if
i * j <= max_num and i * j >= min_num)
target_ratios = sorted(target_ratios, key=lambda x: x[0] * x[1])
# find the closest aspect ratio to the target
target_aspect_ratio = find_closest_aspect_ratio(
aspect_ratio, target_ratios, orig_width, orig_height, image_size)
# calculate the target width and height
target_width = image_size * target_aspect_ratio[0]
target_height = image_size * target_aspect_ratio[1]
blocks = target_aspect_ratio[0] * target_aspect_ratio[1]
# resize the image
resized_img = image.resize((target_width, target_height))
processed_images = []
for i in range(blocks):
box = (
(i % (target_width // image_size)) * image_size,
(i // (target_width // image_size)) * image_size,
((i % (target_width // image_size)) + 1) * image_size,
((i // (target_width // image_size)) + 1) * image_size
)
# split the image
split_img = resized_img.crop(box)
processed_images.append(split_img)
assert len(processed_images) == blocks
if use_thumbnail and len(processed_images) != 1:
thumbnail_img = image.resize((image_size, image_size))
processed_images.append(thumbnail_img)
return processed_images
def load_image(image_file, input_size=448, max_num=12):
image = Image.open(image_file).convert('RGB')
transform = build_transform(input_size=input_size)
images = dynamic_preprocess(image, image_size=input_size, use_thumbnail=True, max_num=max_num)
pixel_values = [transform(image) for image in images]
pixel_values = torch.stack(pixel_values)
return pixel_values
def split_model(model_name):
device_map = {}
world_size = torch.cuda.device_count()
config = AutoConfig.from_pretrained(model_path, trust_remote_code=True)
num_layers = config.llm_config.num_hidden_layers
# Since the first GPU will be used for ViT, treat it as half a GPU.
num_layers_per_gpu = math.ceil(num_layers / (world_size - 0.5))
num_layers_per_gpu = [num_layers_per_gpu] * world_size
num_layers_per_gpu[0] = math.ceil(num_layers_per_gpu[0] * 0.5)
layer_cnt = 0
for i, num_layer in enumerate(num_layers_per_gpu):
for j in range(num_layer):
device_map[f'language_model.model.layers.{layer_cnt}'] = i
layer_cnt += 1
device_map['vision_model'] = 0
device_map['mlp1'] = 0
device_map['language_model.model.tok_embeddings'] = 0
device_map['language_model.model.embed_tokens'] = 0
device_map['language_model.output'] = 0
device_map['language_model.model.norm'] = 0
device_map['language_model.model.rotary_emb'] = 0
device_map['language_model.lm_head'] = 0
device_map[f'language_model.model.layers.{num_layers - 1}'] = 0
return device_map
# If you set `load_in_8bit=True`, you will need two 80GB GPUs.
# If you set `load_in_8bit=False`, you will need at least three 80GB GPUs.
path = 'OpenGVLab/InternVL3-38B'
device_map = split_model('InternVL3-38B')
model = AutoModel.from_pretrained(
path,
torch_dtype=torch.bfloat16,
load_in_8bit=False,
low_cpu_mem_usage=True,
use_flash_attn=True,
trust_remote_code=True,
device_map=device_map).eval()
tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True, use_fast=False)
# set the max number of tiles in `max_num`
pixel_values = load_image('./examples/image1.jpg', max_num=12).to(torch.bfloat16).cuda()
generation_config = dict(max_new_tokens=1024, do_sample=True)
# pure-text conversation (纯文本对话)
question = 'Hello, who are you?'
response, history = model.chat(tokenizer, None, question, generation_config, history=None, return_history=True)
print(f'User: {question}\nAssistant: {response}')
question = 'Can you tell me a story?'
response, history = model.chat(tokenizer, None, question, generation_config, history=history, return_history=True)
print(f'User: {question}\nAssistant: {response}')
# single-image single-round conversation (单图单轮对话)
question = '<image>\nPlease describe the image shortly.'
response = model.chat(tokenizer, pixel_values, question, generation_config)
print(f'User: {question}\nAssistant: {response}')
# single-image multi-round conversation (单图多轮对话)
question = '<image>\nPlease describe the image in detail.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config, history=None, return_history=True)
print(f'User: {question}\nAssistant: {response}')
question = 'Please write a poem according to the image.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config, history=history, return_history=True)
print(f'User: {question}\nAssistant: {response}')
# multi-image multi-round conversation, combined images (多图多轮对话，拼接图像)
pixel_values1 = load_image('./examples/image1.jpg', max_num=12).to(torch.bfloat16).cuda()
pixel_values2 = load_image('./examples/image2.jpg', max_num=12).to(torch.bfloat16).cuda()
pixel_values = torch.cat((pixel_values1, pixel_values2), dim=0)
question = '<image>\nDescribe the two images in detail.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
history=None, return_history=True)
print(f'User: {question}\nAssistant: {response}')
question = 'What are the similarities and differences between these two images.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
history=history, return_history=True)
print(f'User: {question}\nAssistant: {response}')
# multi-image multi-round conversation, separate images (多图多轮对话，独立图像)
pixel_values1 = load_image('./examples/image1.jpg', max_num=12).to(torch.bfloat16).cuda()
pixel_values2 = load_image('./examples/image2.jpg', max_num=12).to(torch.bfloat16).cuda()
pixel_values = torch.cat((pixel_values1, pixel_values2), dim=0)
num_patches_list = [pixel_values1.size(0), pixel_values2.size(0)]
question = 'Image-1: <image>\nImage-2: <image>\nDescribe the two images in detail.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
num_patches_list=num_patches_list,
history=None, return_history=True)
print(f'User: {question}\nAssistant: {response}')
question = 'What are the similarities and differences between these two images.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
num_patches_list=num_patches_list,
history=history, return_history=True)
print(f'User: {question}\nAssistant: {response}')
# batch inference, single image per sample (单图批处理)
pixel_values1 = load_image('./examples/image1.jpg', max_num=12).to(torch.bfloat16).cuda()
pixel_values2 = load_image('./examples/image2.jpg', max_num=12).to(torch.bfloat16).cuda()
num_patches_list = [pixel_values1.size(0), pixel_values2.size(0)]
pixel_values = torch.cat((pixel_values1, pixel_values2), dim=0)
questions = ['<image>\nDescribe the image in detail.'] * len(num_patches_list)
responses = model.batch_chat(tokenizer, pixel_values,
num_patches_list=num_patches_list,
questions=questions,
generation_config=generation_config)
for question, response in zip(questions, responses):
print(f'User: {question}\nAssistant: {response}')
# video multi-round conversation (视频多轮对话)
def get_index(bound, fps, max_frame, first_idx=0, num_segments=32):
if bound:
start, end = bound[0], bound[1]
else:
start, end = -100000, 100000
start_idx = max(first_idx, round(start * fps))
end_idx = min(round(end * fps), max_frame)
seg_size = float(end_idx - start_idx) / num_segments
frame_indices = np.array([
int(start_idx + (seg_size / 2) + np.round(seg_size * idx))
for idx in range(num_segments)
])
return frame_indices
def load_video(video_path, bound=None, input_size=448, max_num=1, num_segments=32):
vr = VideoReader(video_path, ctx=cpu(0), num_threads=1)
max_frame = len(vr) - 1
fps = float(vr.get_avg_fps())
pixel_values_list, num_patches_list = [], []
transform = build_transform(input_size=input_size)
frame_indices = get_index(bound, fps, max_frame, first_idx=0, num_segments=num_segments)
for frame_index in frame_indices:
img = Image.fromarray(vr[frame_index].asnumpy()).convert('RGB')
img = dynamic_preprocess(img, image_size=input_size, use_thumbnail=True, max_num=max_num)
pixel_values = [transform(tile) for tile in img]
pixel_values = torch.stack(pixel_values)
num_patches_list.append(pixel_values.shape[0])
pixel_values_list.append(pixel_values)
pixel_values = torch.cat(pixel_values_list)
return pixel_values, num_patches_list
video_path = './examples/red-panda.mp4'
pixel_values, num_patches_list = load_video(video_path, num_segments=8, max_num=1)
pixel_values = pixel_values.to(torch.bfloat16).cuda()
video_prefix = ''.join([f'Frame{i+1}: <image>\n' for i in range(len(num_patches_list))])
question = video_prefix + 'What is the red panda doing?'
# Frame1: <image>\nFrame2: <image>\n...\nFrame8: <image>\n{question}
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
num_patches_list=num_patches_list, history=None, return_history=True)
print(f'User: {question}\nAssistant: {response}')
question = 'Describe this video in detail.'
response, history = model.chat(tokenizer, pixel_values, question, generation_config,
num_patches_list=num_patches_list, history=history, return_history=True)
print(f'User: {question}\nAssistant: {response}')
```
#### Streaming Output
Besides this method, you can also use the following code to get streamed output.
```python
from transformers import TextIteratorStreamer
from threading import Thread
# Initialize the streamer
streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True, timeout=10)
# Define the generation configuration
generation_config = dict(max_new_tokens=1024, do_sample=False, streamer=streamer)
# Start the model chat in a separate thread
thread = Thread(target=model.chat, kwargs=dict(
tokenizer=tokenizer, pixel_values=pixel_values, question=question,
history=None, return_history=False, generation_config=generation_config,
))
thread.start()
# Initialize an empty string to store the generated text
generated_text = ''
# Loop through the streamer to get the new text as it is generated
for new_text in streamer:
if new_text == model.conv_template.sep:
break
generated_text += new_text
print(new_text, end='', flush=True) # Print each new chunk of generated text on the same line
```
## Finetune
Many repositories now support fine-tuning of the InternVL series models, including [InternVL](https://github.com/OpenGVLab/InternVL), [SWIFT](https://github.com/modelscope/ms-swift), [XTurner](https://github.com/InternLM/xtuner), and others. Please refer to their documentation for more details on fine-tuning.
## Deployment
### LMDeploy
LMDeploy is a toolkit for compressing, deploying, and serving LLMs & VLMs.
```sh
# if lmdeploy<0.7.3, you need to explicitly set chat_template_config=ChatTemplateConfig(model_name='internvl2_5')
pip install lmdeploy>=0.7.3
```
LMDeploy abstracts the complex inference process of multi-modal Vision-Language Models (VLM) into an easy-to-use pipeline, similar to the Large Language Model (LLM) inference pipeline.
#### A 'Hello, world' Example
```python
from lmdeploy import pipeline, TurbomindEngineConfig, ChatTemplateConfig
from lmdeploy.vl import load_image
model = 'OpenGVLab/InternVL3-38B'
image = load_image('https://raw.githubusercontent.com/open-mmlab/mmdeploy/main/tests/data/tiger.jpeg')
pipe = pipeline(model, backend_config=TurbomindEngineConfig(session_len=16384, tp=2), chat_template_config=ChatTemplateConfig(model_name='internvl2_5'))
response = pipe(('describe this image', image))
print(response.text)
```
If `ImportError` occurs while executing this case, please install the required dependency packages as prompted.
#### Multi-images Inference
When dealing with multiple images, you can put them all in one list. Keep in mind that multiple images will lead to a higher number of input tokens, and as a result, the size of the context window typically needs to be increased.
```python
from lmdeploy import pipeline, TurbomindEngineConfig, ChatTemplateConfig
from lmdeploy.vl import load_image
from lmdeploy.vl.constants import IMAGE_TOKEN
model = 'OpenGVLab/InternVL3-38B'
pipe = pipeline(model, backend_config=TurbomindEngineConfig(session_len=16384, tp=2), chat_template_config=ChatTemplateConfig(model_name='internvl2_5'))
image_urls=[
'https://raw.githubusercontent.com/open-mmlab/mmdeploy/main/demo/resources/human-pose.jpg',
'https://raw.githubusercontent.com/open-mmlab/mmdeploy/main/demo/resources/det.jpg'
]
images = [load_image(img_url) for img_url in image_urls]
# Numbering images improves multi-image conversations
response = pipe((f'Image-1: {IMAGE_TOKEN}\nImage-2: {IMAGE_TOKEN}\ndescribe these two images', images))
print(response.text)
```
#### Batch Prompts Inference
Conducting inference with batch prompts is quite straightforward; just place them within a list structure:
```python
from lmdeploy import pipeline, TurbomindEngineConfig, ChatTemplateConfig
from lmdeploy.vl import load_image
model = 'OpenGVLab/InternVL3-38B'
pipe = pipeline(model, backend_config=TurbomindEngineConfig(session_len=16384, tp=2), chat_template_config=ChatTemplateConfig(model_name='internvl2_5'))
image_urls=[
"https://raw.githubusercontent.com/open-mmlab/mmdeploy/main/demo/resources/human-pose.jpg",
"https://raw.githubusercontent.com/open-mmlab/mmdeploy/main/demo/resources/det.jpg"
]
prompts = [('describe this image', load_image(img_url)) for img_url in image_urls]
response = pipe(prompts)
print(response)
```
#### Multi-turn Conversation
There are two ways to do the multi-turn conversations with the pipeline. One is to construct messages according to the format of OpenAI and use above introduced method, the other is to use the `pipeline.chat` interface.
```python
from lmdeploy import pipeline, TurbomindEngineConfig, GenerationConfig, ChatTemplateConfig
from lmdeploy.vl import load_image
model = 'OpenGVLab/InternVL3-38B'
pipe = pipeline(model, backend_config=TurbomindEngineConfig(session_len=16384, tp=2), chat_template_config=ChatTemplateConfig(model_name='internvl2_5'))
image = load_image('https://raw.githubusercontent.com/open-mmlab/mmdeploy/main/demo/resources/human-pose.jpg')
gen_config = GenerationConfig(top_k=40, top_p=0.8, temperature=0.8)
sess = pipe.chat(('describe this image', image), gen_config=gen_config)
print(sess.response.text)
sess = pipe.chat('What is the woman doing?', session=sess, gen_config=gen_config)
print(sess.response.text)
```
#### Service
LMDeploy's `api_server` enables models to be easily packed into services with a single command. The provided RESTful APIs are compatible with OpenAI's interfaces. Below are an example of service startup:
```shell
lmdeploy serve api_server OpenGVLab/InternVL3-38B --chat-template internvl2_5 --server-port 23333 --tp 2
```
To use the OpenAI-style interface, you need to install OpenAI:
```shell
pip install openai
```
Then, use the code below to make the API call:
```python
from openai import OpenAI
client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')
model_name = client.models.list().data[0].id
response = client.chat.completions.create(
model=model_name,
messages=[{
'role':
'user',
'content': [{
'type': 'text',
'text': 'describe this image',
}, {
'type': 'image_url',
'image_url': {
'url':
'https://modelscope.oss-cn-beijing.aliyuncs.com/resource/tiger.jpeg',
},
}],
}],
temperature=0.8,
top_p=0.8)
print(response)
```
## License
This project is released under the MIT License. This project uses the pre-trained Qwen2.5 as a component, which is licensed under the Qwen License.
## Citation
If you find this project useful in your research, please consider citing:
```BibTeX
@article{chen2024expanding,
title={Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling},
author={Chen, Zhe and Wang, Weiyun and Cao, Yue and Liu, Yangzhou and Gao, Zhangwei and Cui, Erfei and Zhu, Jinguo and Ye, Shenglong and Tian, Hao and Liu, Zhaoyang and others},
journal={arXiv preprint arXiv:2412.05271},
year={2024}
}
@article{wang2024mpo,
title={Enhancing the Reasoning Ability of Multimodal Large Language Models via Mixed Preference Optimization},
author={Wang, Weiyun and Chen, Zhe and Wang, Wenhai and Cao, Yue and Liu, Yangzhou and Gao, Zhangwei and Zhu, Jinguo and Zhu, Xizhou and Lu, Lewei and Qiao, Yu and Dai, Jifeng},
journal={arXiv preprint arXiv:2411.10442},
year={2024}
}
@article{chen2024far,
title={How Far Are We to GPT-4V? Closing the Gap to Commercial Multimodal Models with Open-Source Suites},
author={Chen, Zhe and Wang, Weiyun and Tian, Hao and Ye, Shenglong and Gao, Zhangwei and Cui, Erfei and Tong, Wenwen and Hu, Kongzhi and Luo, Jiapeng and Ma, Zheng and others},
journal={arXiv preprint arXiv:2404.16821},
year={2024}
}
@inproceedings{chen2024internvl,
title={Internvl: Scaling up vision foundation models and aligning for generic visual-linguistic tasks},
author={Chen, Zhe and Wu, Jiannan and Wang, Wenhai and Su, Weijie and Chen, Guo and Xing, Sen and Zhong, Muyan and Zhang, Qinglong and Zhu, Xizhou and Lu, Lewei and others},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={24185--24198},
year={2024}
}
``` |
liserman/manifestoberta_domains_easyread_german_context_2025-5-20 | liserman | 2025-05-29T11:08:22Z | 8 | 0 | transformers | [
"transformers",
"safetensors",
"xlm-roberta",
"text-classification",
"custom_code",
"arxiv:1910.09700",
"autotrain_compatible",
"region:us"
] | text-classification | 2025-05-21T10:01:45Z | ---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
mlabonne/gemma-3-1b-it-abliterated-v2-GGUF | mlabonne | 2025-05-29T10:59:07Z | 37 | 1 | transformers | [
"transformers",
"gguf",
"image-text-to-text",
"license:gemma",
"endpoints_compatible",
"region:us",
"conversational"
] | image-text-to-text | 2025-05-28T16:27:27Z | ---
license: gemma
library_name: transformers
pipeline_tag: image-text-to-text
base_model: mlabonne/gemma3-1b-it-abliterated-v2
---
# 💎 Gemma 3 1B IT Abliterated GGUF
<center>Gemma 3 Abliterated GGUF <a href="https://huggingface.co/mlabonne/gemma-3-1b-it-abliterated-v2-GGUF">1B</a> • <a href="https://huggingface.co/mlabonne/gemma-3-4b-it-abliterated-v2-GGUF">4B</a> • <a href="https://huggingface.co/mlabonne/gemma-3-12b-it-abliterated-v2-GGUF">12B</a> • <a href="https://huggingface.co/mlabonne/gemma-3-27b-it-abliterated-v2-GGUF">27B</a></center>
This is an uncensored version of [google/gemma-3-1b-it](https://huggingface.co/google/gemma-3-1b-it) created with a new abliteration technique.
See [this article](https://huggingface.co/blog/mlabonne/abliteration) to know more about abliteration.
This is a new, improved version that targets refusals with enhanced accuracy.
I recommend using these generation parameters: `temperature=1.0`, `top_k=64`, `top_p=0.95`.
## ⚡️ Quantization
* **GGUF**: https://huggingface.co/mlabonne/gemma-3-1b-it-abliterated-v2-GGUF
## ✂️ Abliteration
The refusal direction is computed by comparing the residual streams between target (harmful) and baseline (harmless) samples.
The hidden states of target modules (e.g., o_proj) are orthogonalized to subtract this refusal direction with a given weight factor.
These weight factors follow a normal distribution with a certain spread and peak layer.
Modules can be iteratively orthogonalized in batches, or the refusal direction can be accumulated to save memory.
Finally, I used a hybrid evaluation with a dedicated test set to calculate the acceptance rate. This uses both a dictionary approach and [NousResearch/Minos-v1](https://huggingface.co/NousResearch/Minos-v1).
The goal is to obtain an acceptance rate >90% and still produce coherent outputs. |
BootesVoid/cmb97xxhf08gi1b1ygoutv7x7_cmb9861oa08ma1b1y3lwfx1yn | BootesVoid | 2025-05-29T10:47:44Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T10:47:43Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: emma_grace
---
# Cmb97Xxhf08Gi1B1Ygoutv7X7_Cmb9861Oa08Ma1B1Y3Lwfx1Yn
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `emma_grace` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "emma_grace",
"lora_weights": "https://huggingface.co/BootesVoid/cmb97xxhf08gi1b1ygoutv7x7_cmb9861oa08ma1b1y3lwfx1yn/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('BootesVoid/cmb97xxhf08gi1b1ygoutv7x7_cmb9861oa08ma1b1y3lwfx1yn', weight_name='lora.safetensors')
image = pipeline('emma_grace').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 2000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/BootesVoid/cmb97xxhf08gi1b1ygoutv7x7_cmb9861oa08ma1b1y3lwfx1yn/discussions) to add images that show off what you’ve made with this LoRA.
|
mlabonne/gemma-3-27b-it-qat-abliterated | mlabonne | 2025-05-29T10:46:44Z | 1 | 3 | transformers | [
"transformers",
"safetensors",
"gemma3_text",
"text-generation",
"image-text-to-text",
"conversational",
"base_model:google/gemma-3-27b-it-qat-q4_0-unquantized",
"base_model:finetune:google/gemma-3-27b-it-qat-q4_0-unquantized",
"license:gemma",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | image-text-to-text | 2025-05-28T21:43:21Z | ---
license: gemma
library_name: transformers
pipeline_tag: image-text-to-text
base_model: google/gemma-3-27b-it-qat-q4_0-unquantized
---
# 💎 Gemma 3 27B IT QAT Abliterated
<center>Gemma 3 QAT Abliterated <a href="https://huggingface.co/mlabonne/gemma-3-1b-it-qat-abliterated">1B</a> • <a href="https://huggingface.co/mlabonne/gemma-3-4b-it-qat-abliterated">4B</a> • <a href="https://huggingface.co/mlabonne/gemma-3-12b-it-qat-abliterated">12B</a> • <a href="https://huggingface.co/mlabonne/gemma-3-27b-it-qat-abliterated">27B</a></center>
This is an uncensored version of [google/gemma-3-27b-it-qat-q4_0-unquantized](https://huggingface.co/google/gemma-3-27b-it-qat-q4_0-unquantized) created with a new abliteration technique.
See [this article](https://huggingface.co/blog/mlabonne/abliteration) to know more about abliteration.
This is a new, improved version that targets refusals with enhanced accuracy.
I recommend using these generation parameters: `temperature=1.0`, `top_k=64`, `top_p=0.95`.
## ✂️ Abliteration
The refusal direction is computed by comparing the residual streams between target (harmful) and baseline (harmless) samples.
The hidden states of target modules (e.g., o_proj) are orthogonalized to subtract this refusal direction with a given weight factor.
These weight factors follow a normal distribution with a certain spread and peak layer.
Modules can be iteratively orthogonalized in batches, or the refusal direction can be accumulated to save memory.
Finally, I used a hybrid evaluation with a dedicated test set to calculate the acceptance rate. This uses both a dictionary approach and [NousResearch/Minos-v1](https://huggingface.co/NousResearch/Minos-v1).
The goal is to obtain an acceptance rate >90% and still produce coherent outputs. |
Varinder2110/23a291cc-7a62-44f4-8e35-8a85f61f6fba | Varinder2110 | 2025-05-29T10:45:05Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T09:39:52Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: TOK
---
# 23A291Cc 7A62 44F4 8E35 8A85F61F6Fba
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `TOK` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "TOK",
"lora_weights": "https://huggingface.co/Varinder2110/23a291cc-7a62-44f4-8e35-8a85f61f6fba/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('Varinder2110/23a291cc-7a62-44f4-8e35-8a85f61f6fba', weight_name='lora.safetensors')
image = pipeline('TOK').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 6000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/Varinder2110/23a291cc-7a62-44f4-8e35-8a85f61f6fba/discussions) to add images that show off what you’ve made with this LoRA.
|
snezhanata/advertisment_instraction_mistral-test-dev | snezhanata | 2025-05-29T10:43:59Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"mistral",
"text-generation",
"llama-factory",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-29T09:22:05Z | ---
library_name: transformers
tags:
- llama-factory
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
batmangiaicuuthegioi/NODUPLICATES_512_10000 | batmangiaicuuthegioi | 2025-05-29T10:34:11Z | 0 | 0 | sentence-transformers | [
"sentence-transformers",
"safetensors",
"xlm-roberta",
"sentence-similarity",
"feature-extraction",
"generated_from_trainer",
"dataset_size:36728",
"loss:MultipleNegativesRankingLoss",
"dataset:batmangiaicuuthegioi/augmentated_legal_triplets",
"arxiv:1908.10084",
"arxiv:1705.00652",
"base_model:BAAI/bge-m3",
"base_model:finetune:BAAI/bge-m3",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | sentence-similarity | 2025-05-29T10:33:28Z | ---
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- generated_from_trainer
- dataset_size:36728
- loss:MultipleNegativesRankingLoss
base_model: BAAI/bge-m3
widget:
- source_sentence: Chức năng bảo hộ lãnh sự của lãnh sự danh dự được quy định thế
nào?
sentences:
- Điều 121. Bảo lĩnh. khoản 1. bảo lĩnh là biện pháp ngăn chặn thay thế tạm giam.
căn cứ vào tính chất, mức độ nguy hiểm cho xã hội của hành vi và nhân thân của
bị can, bị cáo, cơ quan điều tra, viện kiểm sát, tòa án có thể quyết định cho
họ được bảo lĩnh. Bảo lĩnh. khoản 2. cơ quan, tổ chức có thể nhận bảo lĩnh cho
bị can, bị cáo là người của cơ quan, tổ chức mình. cơ quan, tổ chức nhận bảo lĩnh
phải có giấy cam đoan và có xác nhận của người đứng đầu cơ quan, tổ chức. cá nhân
là người đủ 18 tuổi trở lên, nhân thân tốt, nghiêm chỉnh chấp hành pháp luật,
thu nhập ổn định và có điều kiện quản lý người được bảo lĩnh thì có thể nhận bảo
lĩnh cho bị can, bị
- Điều 8. Chức năng bảo hộ lãnh sự. khoản 1. lãnh sự danh dự áp dụng mọi biện pháp
thích hợp để bảo hộ lãnh sự đối với lợi ích của nhà nước, quyền và lợi ích hợp
pháp của công dân, pháp nhân việt nam trong khu vực lãnh sự phù hợp với pháp luật
việt nam, pháp luật của nước tiếp nhận và điều ước quốc tế mà việt nam và nước
tiếp nhận là thành viên, phù hợp với pháp luật và thông lệ quốc tế. Chức năng
bảo hộ lãnh sự. khoản 2. lãnh sự danh dự có nghĩa vụ thông báo về mọi thông tin
có được trong trường hợp công dân việt nam bị bắt, tạm giữ, tạm giam hoặc chấp
hành hình phạt tù trong khu vực lãnh sự cho cơ quan đại diện ngoại giao việt nam
tại nước tiếp nhận hoặc
- 'Theo quy định tại Khoản 2 Điều 18 Luật Bảo hiểm xã hội 2014 thì người lao động
có quyền "Được cấp và quản lý sổ bảo hiểm xã hội". Khi Anh/Chị nghỉ việc, thì
theo quy định tại Khoản 3 Điều 47 Bộ luật Lao động 2012: “Người sử dụng lao động
có trách nhiệm hoàn thành thủ tục xác nhận và trả lại sổ bảo hiểm xã hội và những
giấy tờ khác mà người sử dụng lao động đã giữ lại của người lao động”. Trường
hợp công ty Anh/Chị không thực hiện đúng theo quy định của pháp luật, Anh/Chị
có thể thực hiện việc khiếu nại hành vi sai phạm này theo quy định tại Luật Khiếu
nại. Trường hợp công ty không giải quyết thoả đáng, Anh/Chị có thể khiếu nại trực
tiếp đến Thanh tra lao động thuộc Sở Lao động - Thương binh và Xã hội hoặc thực
hiện việc khởi kiện trực tiếp ra Toà án theo quy định tại Điểm d Khoản 1 Điều
201 Bộ Luật lao động và Điểm d Khoản 1 Điều 32 Bộ Luật Tố tụng dân sự 2015: "Điều
201. Trình tự, thủ tục hoà giải tranh chấp lao động cá nhân của hoà giải viên
lao động 1. Tranh chấp lao động cá nhân phải thông qua thủ tục hoà giải của hoà
giải viên lao động trước khi yêu cầu toà án giải quyết, trừ các tranh chấp lao
động sau đây không bắt buộc phải qua thủ tục hoà giải: . .. d) Về bảo hiểm xã
hội theo quy định của pháp luật về bảo hiểm xã hội, về bảo hiểm y tế theo quy
định của pháp luật về bảo hiểm y tế. " "Điều 32. Những tranh chấp về lao động
thuộc thẩm quyền giải quyết của Toà án 1. Tranh chấp lao động cá nhân giữa người
lao động với người sử dụng lao động phải thông qua thủ tục hoà giải của hoà giải
viên lao động mà hoà giải thành nhưng các bên không thực hiện hoặc thực hiện không
đúng, hoà giải không thành hoặc không hoà giải trong thời hạn do pháp luật quy
định, trừ các tranh chấp lao động sau đây không bắt buộc phải qua thủ tục hoà
giải: . .. d) Về bảo hiểm xã hội theo quy định của pháp luật về bảo hiểm xã hội,
về bảo hiểm y tế theo quy định của pháp luật về bảo hiểm y tế, về bảo hiểm thất
nghiệp theo quy định của pháp luật về việc làm, về bảo hiểm tai nạn lao động,
bệnh nghề nghiệp theo quy định của pháp luật về an toàn, vệ sinh lao động; " Ban
biên tập thông tin đến Anh/Chị! Trân trọng!'
- source_sentence: Theo Thông tư mới nhất về giao dịch cổ phiếu niêm yết, đăng ký
giao dịch chứng khoán thì quy chế giao dịch chứng khoán do ai ban hành?
sentences:
- 'Khoản 3 Điều 3 Thông tư 120/2020/TT-BTC (Có hiệu lực từ 15/02/2021) quy định
về tổ chức giao dịch chứng khoán như sau: Sở giao dịch chứng khoán Việt Nam ban
hành quy chế giao dịch chứng khoán bao gồm các nội dung cơ bản sau: phương thức
giao dịch; thời gian giao dịch; cách xác định giá tham chiếu; biên độ dao động
giá chứng khoán; cơ chế ngắt mạch thị trường (nếu có) ; các loại lệnh giao dịch;
việc sửa lệnh, huỷ lệnh giao dịch; việc xác lập giao dịch và loại bỏ giao dịch
chứng khoán; việc tạm ngừng giao dịch, đình chỉ một phần hoặc toàn bộ giao dịch
của một mã chứng khoán; việc công bố thông tin về kết quả giao dịch và các nội
dung khác có liên quan. Như vậy, quy chế giao dịch chứng khoán do Sở giao dịch
chứng khoán Việt Nam ban hành. Trân trọng!'
- Điều 6. Xử phạt người điều khiển xe mô tô, xe gắn máy (kể cả xe máy điện), các
loại xe tương tự xe mô tô và các loại xe tương tự xe gắn máy vi phạm quy tắc giao
thông đường bộ. điểm q) điều khiển xe chạy dưới tốc độ tối thiểu trên những đoạn
đường bộ có quy định tốc độ tối thiểu cho phép. Xử phạt người điều khiển xe mô
tô, xe gắn máy (kể cả xe máy điện), các loại xe tương tự xe mô tô và các loại
xe tương tự xe gắn máy vi phạm quy tắc giao thông đường bộ. điểm m) ngồi phía
sau vòng tay qua người ngồi trước để điều khiển xe, trừ trường hợp chở trẻ em
ngồi phía trước. Xử phạt người điều khiển xe mô tô, xe gắn máy (kể cả xe máy điện),
các
- 'Quyền của Sở giao dịch chứng khoán Việt Nam? Xin chào anh/chị, em đang tìm hiểu
về quy định mới điều chỉnh về hoạt động chứng khoán, cho em hỏi theo quy định
mới thì Sở giao dịch chứng khoán Việt Nam có những quyền nào? Trả lời: Căn cứ
Khoản 1 Điều 46 Luật chứng khoán 2019 quy định Sở giao dịch chứng khoán Việt Nam
có các quyền sau đây: - Ban hành các quy chế về niêm yết chứng khoán, giao dịch
chứng khoán, công bố thông tin, thành viên của Sở giao dịch chứng khoán Việt Nam
và các quy chế nghiệp vụ khác liên quan đến tổ chức và hoạt động thị trường giao
dịch chứng khoán sau khi được Ủy ban Chứng khoán Nhà nước chấp thuận; - Tổ chức,
vận hành thị trường giao dịch chứng khoán; - Cảnh báo, kiểm soát, hạn chế giao
dịch chứng khoán theo quy định của pháp luật và quy chế của Sở giao dịch chứng
khoán Việt Nam; - Tạm ngừng, đình chỉ giao dịch đối với một hoặc một số chứng
khoán trong trường hợp giá, khối lượng giao dịch chứng khoán có biến động bất
thường, tổ chức niêm yết, tổ chức đăng ký giao dịch không có biện pháp khắc phục
nguyên nhân dẫn đến việc chứng khoán bị đưa vào diện cảnh báo, kiểm soát, hạn
chế giao dịch hoặc trong trường hợp cần thiết để bảo vệ quyền, lợi ích hợp pháp
của nhà đầu tư và bảo đảm ổn định, an toàn của thị trường chứng khoán; - Chấp
thuận, thay đổi, huỷ bỏ niêm yết, đăng ký giao dịch chứng khoán và giám sát việc
duy trì điều kiện niêm yết chứng khoán của các tổ chức niêm yết; - Chấp thuận,
huỷ bỏ tư cách thành viên của Sở giao dịch chứng khoán Việt Nam; - Cung cấp dịch
vụ đấu giá, đấu thầu; dịch vụ về thông tin thị trường và thông tin liên quan đến
chứng khoán niêm yết, đăng ký giao dịch; dịch vụ phát triển hạ tầng công nghệ
cho thị trường chứng khoán và các dịch vụ liên quan khác theo quy định tại Điều
lệ Sở giao dịch chứng khoán Việt Nam; - Làm trung gian hoà giải theo yêu cầu của
thành viên của Sở giao dịch chứng khoán Việt Nam khi phát sinh tranh chấp liên
quan đến hoạt động giao dịch chứng khoán; - Kiểm tra, xử lý vi phạm đối với thành
viên của Sở giao dịch chứng khoán Việt Nam, tổ chức niêm yết, tổ chức đăng ký
giao dịch theo quy chế của Sở giao dịch chứng khoán Việt Nam; - Đề nghị cơ quan
quản lý nhà nước cung cấp thông tin liên quan đến thành viên của Sở giao dịch
chứng khoán Việt Nam, tổ chức niêm yết, tổ chức đăng ký giao dịch để phục vụ công
bố thông tin theo quy định của pháp luật; - Quyền khác theo quy định của pháp
luật'
- source_sentence: Năm 2006, tôi kết hôn nên đã nhập hộ khẩu vào gia đình nhà chồng
ở Bình Dương. Tuy nhiên, hiện nay tôi và chồng đã ly dị, tôi muốn xin tách khẩu
khỏi gia đình nhà chồng và nhập lại hộ khẩu vào gia đình nhà tôi (Nam Định). Luật
sư cho tôi hỏi, như vậy có được không ạ? Thời hạn giải quyết là bao lâu?
sentences:
- 'Theo thông tin bạn cung cấp thì chúng tôi hiểu rằng vợ chồng em bạn đã ly hôn
4 năm. Hiện nay em bạn muốn ly hôn nhưng không biết địa chỉ của chồng. Căn cứ
tại Điểm a Khoản 1 Điều 40 Bộ luật Tố tụng Dân sự 2015 quy định như sau: Nguyên
đơn có quyền lựa chọn Toà án giải quyết tranh chấp về dân sự, hôn nhân và gia
đình, kinh doanh, thương mại, lao động trong các trường hợp sau đây: a) Nếu không
biết nơi cư trú, làm việc, trụ sở của bị đơn thì nguyên đơn có thể yêu cầu Toà
án nơi bị đơn cư trú, làm việc, có trụ sở cuối cùng hoặc nơi bị đơn có tài sản
giải quyết; b) Nếu tranh chấp phát sinh từ hoạt động của chi nhánh tổ chức thì
nguyên đơn có thể yêu cầu Toà án nơi tổ chức có trụ sở hoặc nơi tổ chức có chi
nhánh giải quyết; c) Nếu bị đơn không có nơi cư trú, làm việc, trụ sở ở Việt Nam
hoặc vụ án về tranh chấp việc cấp dưỡng thì nguyên đơn có thể yêu cầu Toà án nơi
mình cư trú, làm việc, có trụ sở giải quyết; d) Nếu tranh chấp về bồi thường thiệt
hại ngoài hợp đồng thì nguyên đơn có thể yêu cầu Toà án nơi mình cư trú, làm việc,
có trụ sở hoặc nơi xảy ra việc gây thiệt hại giải quyết; đ) Nếu tranh chấp về
bồi thường thiệt hại, trợ cấp khi chấm dứt hợp đồng lao động, bảo hiểm xã hội,
bảo hiểm y tế, bảo hiểm thất nghiệp, quyền và lợi ích liên quan đến việc làm,
tiền lương, thu nhập và các điều kiện lao động khác đối với người lao động thì
nguyên đơn là người lao động có thể yêu cầu Toà án nơi mình cư trú, làm việc giải
quyết; e) Nếu tranh chấp phát sinh từ việc sử dụng lao động của người cai thầu
hoặc người có vai trò trung gian thì nguyên đơn có thể yêu cầu Toà án nơi người
sử dụng lao động là chủ chính cư trú, làm việc, có trụ sở hoặc nơi người cai thầu,
người có vai trò trung gian cư trú, làm việc giải quyết; g) Nếu tranh chấp phát
sinh từ quan hệ hợp đồng thì nguyên đơn có thể yêu cầu Toà án nơi hợp đồng được
thực hiện giải quyết; h) Nếu các bị đơn cư trú, làm việc, có trụ sở ở nhiều nơi
khác nhau thì nguyên đơn có thể yêu cầu Toà án nơi một trong các bị đơn cư trú,
làm việc, có trụ sở giải quyết; i) Nếu tranh chấp bất động sản mà bất động sản
có ở nhiều địa phương khác nhau thì nguyên đơn có thể yêu cầu Toà án nơi có một
trong các bất động sản giải quyết. Nếu hiện tại em bạn muốn ly hôn thì em bạn
phải'
- 'Người nhận con nuôi phải đáp ứng điều kiện gì? Theo Điều 14 Luật Nuôi con nuôi
2010 quy định về người nhận con nuôi phải đáp ứng điều kiện sau đây: - Người nhận
con nuôi phải có đủ các điều kiện sau đây: + Có năng lực hành vi dân sự đầy đủ;
+ Hơn con nuôi từ 20 tuổi trở lên; + Có điều kiện về sức khoẻ, kinh tế, chỗ ở
bảo đảm việc chăm sóc, nuôi dưỡng, giáo dục con nuôi; + Có tư cách đạo đức tốt.
- Những người sau đây không được nhận con nuôi: + Đang bị hạn chế một số quyền
của cha, mẹ đối với con chưa thành niên; + Đang chấp hành quyết định xử lý hành
chính tại cơ sở giáo dục, cơ sở chữa bệnh; + Đang chấp hành hình phạt tù; + Chưa
được xoá án tích về một trong các tội cố ý xâm phạm tính mạng, sức khoẻ, nhân
phẩm, danh dự của người khác; + Ngược đãi hoặc hành hạ ông bà, cha mẹ, vợ chồng,
con, cháu, người có công nuôi dưỡng mình; + Dụ dỗ, ép buộc hoặc chứa chấp người
chưa thành niên vi phạm pháp luật; mua bán, đánh tráo, chiếm đoạt trẻ em. - Trường
hợp cha dượng nhận con riêng của vợ, mẹ kế nhận con riêng của chồng làm con nuôi
hoặc cô, cậu, dì, chú, bác ruột nhận cháu làm con nuôi thì không áp dụng quy định
tại điểm b và điểm c khoản 1 Điều 14 Luật Nuôi con nuôi 2010. Con nuôi sẽ được
thừa kế tài sản hợp pháp trong trường hợp nào? (Hình từ Internet) Người nhận nuôi
con nuôi phải chuẩn bị những hồ sơ gì? Căn cứ theo Điều 17 Luật Nuôi con nuôi
2010 quy định về hồ sơ của người nhận con nuôi cụ thể như sau: Hồ sơ của người
nhận con nuôi Hồ sơ của người nhận con nuôi gồm có: 1. Đơn xin nhận con nuôi;
2. Bản sao Hộ chiếu, Giấy chứng minh nhân dân hoặc giấy tờ có giá trị thay thế;
3. Phiếu lý lịch tư pháp; 4. Văn bản xác nhận tình trạng hôn nhân; 5. Giấy khám
sức khoẻ do cơ quan y tế cấp huyện trở lên cấp; văn bản xác nhận hoàn cảnh gia
đình, tình trạng chỗ ở, điều kiện kinh tế do Ủy ban nhân dân cấp xã nơi người
nhận con nuôi thường trú cấp, trừ trường hợp quy định tại khoản 3 Điều 14 của
Luật này. Theo đó, người nhận nuôi con nuôi phải chuẩn bị những hồ sơ sau đây:
- Đơn xin nhận con nuôi; - Bản sao Hộ chiếu, Giấy chứng minh nhân dân hoặc giấy
tờ có giá trị thay thế; - Phiếu lý lịch tư pháp; - Văn bản xác nhận tình trạng
hôn nhân; - Giấy khám sức khoẻ do cơ quan y tế cấp huyện trở lên cấp; - Văn bản
xác nhận hoàn cảnh'
- 'Theo quy định Luật cư trú trường hợp có cùng chỗ ở hợp pháp được tách sổ hộ khẩu
bao gồm: - Người có năng lực hành vi dân sự đầy đủ và có nhu cầu tách sổ hộ khẩu;
- Người đã nhập vào sổ hộ khẩu quy định tại khoản 3 Điều 25 và khoản 2 Điều 26
của Luật này mà được chủ hộ đồng ý cho tách sổ hộ khẩu bằng văn bản. Theo thông
tin bạn cung cấp, bạn lấy chồng và nhập khẩu vào gia đình nhà chồng, bạn có chỗ
ở hợp pháp và được sự cho phép của chủ hộ cho phép nhập hộ khẩu vào gia đình họ
nên bạn và gia đình có quyền tách sổ hộ khẩu theo điểm b Khoản 1 Điều 27 trên.
Tuy nhiên trong trường hợp của bạn muốn tách sổ hộ khẩu cần có sự đồng ý của chủ
hộ bằng văn bản và người đến làm thủ tục phải xuất trình sổ hộ khẩu; phiếu báo
thay đổi hộ khẩu, nhân khẩu; ý kiến đồng ý bằng văn bản của chủ hộ cho cơ quan
nhà nước có thẩm quyền. Trong thời hạn bảy ngày làm việc, kể từ ngày nhận đủ hồ
sơ, cơ quan có thẩm quyền phải trả kết quả giải quyết việc tách hộ khẩu; trường
hợp không giải quyết việc tách sổ hộ khẩu thì phải trả lời bằng văn bản và nêu
rõ lý do. Đối với trường hợp bạn muốn chuyển khẩu về gia đình nhà bố mẹ đẻ: Theo
quy định tại Điều 19 Luật cư trú về điều kiện đăng ký thường trú tại tỉnh: Công
dân có chỗ ở hợp pháp ở tỉnh nào thì được đăng ký thường trú tại tỉnh đó. Trường
hợp chỗ ở hợp pháp do thuê, mượn, ở nhờ của cá nhân thì phải được người cho thuê,
cho mượn, cho ở nhờ đồng ý bằng văn bản. Như vậy, khi bạn có chỗ ở hợp pháp ở
Nam Định thì được đăng ký thường trú tại tỉnh đó. Để đăng ký thường trú tại Nam
Định, bạn cần thực hiện theo trình tự thủ tục quy định tại Điều 21 Luật cư trú
và được sự đồng ý cho nhập vào sổ hộ khẩu của chủ hộ nơi đến (bố, mẹ đẻ của bạn)
. Trong thời hạn mười lăm ngày, kể từ ngày nhận đủ hồ sơ, cơ quan có thẩm quyền
quy định tại khoản 1 Điều này phải cấp sổ hộ khẩu cho người đã nộp hồ sơ đăng
ký thường trú; trường hợp không cấp phải trả lời bằng văn bản và nêu rõ lý do.'
- source_sentence: Lệ phí cấp đổi giấy chứng nhận hành nghề chứng khoán cho cá nhân
hành nghề chứng khoán là bao nhiêu? Xin chào Ban biên tập, tôi là Kim Yến hiện
đang sống và làm việc tại Đồng Nai. Tôi hiện đang tìm hiểu về phí và lệ phí trong
lĩnh vực chứng khoán. Vậy Ban biên tập cho tôi hỏi lệ phí cấp đổi giấy chứng nhận
hành nghề chứng khoán cho cá nhân hành nghề chứng khoán là bao nhiêu? Vấn đề này
được quy định cụ thể tại văn bản nào?
sentences:
- 'Cấp lại chứng chỉ hành nghề chứng khoán trong những trường hợp nào? Những trường
hợp nào được cấp lại chứng chỉ hành nghề chứng khoán theo quy định hiện hành?
Trả lời: Căn cứ Khoản 1 Điều 214 Nghị định 155/2020/NĐ-CP quy định về nội dung
trên như sau: 1. Trường hợp được cấp lại chứng chỉ hành nghề chứng khoán. a) Chứng
chỉ hành nghề chứng khoán bị thu hồi theo quy định tại điểm a, c khoản 3 Điều
97 Luật Chứng khoán hoặc bị hỏng, bị mất; b) Thông tin xác nhận nhân thân của
người hành nghề ghi trong chứng chỉ hành nghề chứng khoán thay đổi (số giấy chứng
minh nhân dân hoặc căn cước công dân hoặc số hộ chiếu, quốc tịch, họ tên, ngày
tháng năm sinh) . Nguyên tắc hành nghề chứng khoán của người được cấp CCHN chứng
khoán Nguyên tắc hành nghề chứng khoán của người được cấp chứng chỉ hành nghề
chứng khoán và tổ chức sử dụng người hành nghề chứng khoán được quy định thế nào?
Trả lời: Căn cứ Khoản 2 Điều 216 Nghị định 155/2020/NĐ-CP quy định về về nguyên
tắc hành nghề chứng khoán như sau: - Người có chứng chỉ hành nghề môi giới chứng
khoán được thực hiện nghiệp vụ môi giới chứng khoán, tư vấn đầu tư chứng khoán;
- Người có chứng chỉ hành nghề phân tích tài chính được thực hiện nghiệp vụ môi
giới chứng khoán, tư vấn đầu tư chứng khoán, tự doanh chứng khoán, bảo lãnh phát
hành chứng khoán; - Người có chứng chỉ hành nghề quản lý quỹ được thực hiện nghiệp
vụ môi giới chứng khoán, tư vấn đầu tư chứng khoán, tự doanh chứng khoán, bảo
lãnh phát hành chứng khoán, quản lý danh mục đầu tư chứng khoán, quản lý quỹ đầu
tư chứng khoán; - Chứng chỉ hành nghề chứng khoán chỉ có giá trị sử dụng khi người
được cấp chứng chỉ làm việc tại một công ty chứng khoán, công ty quản lý quỹ đầu
tư chứng khoán, chi nhánh công ty chứng khoán, công ty quản lý quỹ nước ngoài
tại Việt Nam, công ty đầu tư chứng khoán và được công ty đó thông báo với Ủy ban
Chứng khoán Nhà nước; - Người có 01 trong 03 loại chứng chỉ hành nghề chứng khoán
theo quy định tại điểm a, b, c khoản này và có chứng chỉ chuyên môn chứng khoán
phái sinh và thị trường chứng khoán phái sinh được thực hiện nghiệp vụ tương ứng
với chứng chỉ đang nắm giữ liên quan đến chứng khoán phái sinh tại công ty chứng
khoán, công ty quản lý quỹ đầu tư chứng khoán; - Người có chứng chỉ hành nghề
chứng khoán chỉ được làm việc tại 01 bộ phận nghiệp vụ kinh doanh chứng khoán
trong một thời điểm. Cấp lại chứng chỉ hành nghề chứng khoán phải chuẩn bị những
giấy tờ gì? Cấp lại chứng chỉ hành nghề chứng khoán sẽ'
- 'Căn cứ Khoản 12 Điều 22 Nghị định 10/2020/NĐ-CP (Có hiệu lực ngày 01/4/2020)
quy định trách nhiệm của đơn vị kinh doanh vận tải như sau: - Phải nộp lại phù
hiệu, biển hiệu cho Sở Giao thông vận tải khi nhận được quyết định thu hồi phù
hiệu, biển hiệu. - Không được sử dụng xe ô tô để kinh doanh vận tải trong thời
gian xe ô tô bị cơ quan có thẩm quyền áp dụng hình thức xử phạt tước quyền sử
dụng (Giấy phép kinh doanh vận tải bằng xe ô tô, phù hiệu, biển hiệu) hoặc bị
thu hồi phù hiệu, biển hiệu. Chúng tôi phản hồi thông tin đến bạn! Trân trọng!'
- 'Ngày 14/11/2016, Bộ trưởng Bộ Tài chính ban hành Thông tư 272/2016/TT-BTC quy
định mức thu, chế độ thu, nộp, quản lý và sử dụng phí, lệ phí trong lĩnh vực chứng
khoán. Thông tư này quy định mức thu, chế độ thu, nộp, quản lý và sử dụng phí
quản lý và giám sát hoạt động chứng khoán; lệ phí cấp giấy phép, giấy chứng nhận
hoạt động trong lĩnh vực chứng khoán áp dụng tại Ủy ban Chứng khoán Nhà nước.
Thông tư này áp dụng đối với tổ chức, cá nhân nộp phí, lệ phí; tổ chức thu phí,
lệ phí; các tổ chức, cá nhân khác liên quan đến thu, nộp phí, lệ phí. Lệ phí cấp
đổi giấy chứng nhận hành nghề chứng khoán cho cá nhân hành nghề chứng khoán được
quy định tại Biểu mức thu phí, lệ phí ban hành kèm theo Thông tư 272/2016/TT-BTC,
cụ thể: Lệ phí cấp đổi giấy chứng nhận hành nghề chứng khoán cho cá nhân hành
nghề chứng khoán là 01 triệu đồng/lần cấp. Trên đây là tư vấn về lệ phí cấp đổi
giấy chứng nhận hành nghề chứng khoán cho cá nhân hành nghề chứng khoán. Để biết
thêm thông tin chi tiết bạn nên tham khảo tại Thông tư 272/2016/TT-BTC. Mong rằng
những tư vấn của chúng tôi sẽ giúp giải đáp được những vướng mắc của bạn. Chào
thân ái và chúc sức khoẻ!'
- source_sentence: 'Chào Ban biên tập, tôi là Minh Tùng, hiện tôi đang làm việc tại
một cơ quan hành chính nhà nước. Tôi có thắc mắc sau mong sớm nhận phản hồi. Cụ
thể: Sổ bảo hiểm xã hội được cấp lại trong những trường hợp nào?'
sentences:
- 'Theo quy định tại Điều 24 Quy chế hoạt động đăng ký và chuyển quyền sở hữu
chứng khoán do Trung tâm Lưu ký Chứng khoán Việt Nam ban hành kèm theo Quyết
định 196/QĐ-VSD năm 2017 thì: - Đối với các chứng khoán niêm yết, đăng ký giao
dịch thực hiện giao dịch mua, bán qua hệ thống giao dịch của các SGDCK, VSD chuyển
quyền sở hữu căn cứ vào kết quả giao dịch mua, bán của nhà đầu tư do SGDCK cung
cấp. - Đối với các chứng khoán niêm yết, đăng ký giao dịch nhưng không thể thực
hiện giao dịch mua, bán qua hệ thống giao dịch của SGDCK hoặc các giao dịch không
mang tính chất mua bán, VSD chỉ thực hiện chuyển quyền sở hữu đối với các trường
hợp quy định tại Điểm b Khoản 2 Điều 19 Thông tư 05/2015/TT-BTC ngày 15/01/2015
của Bộ trưởng Bộ Tài chính hướng dẫn về hoạt động đăng ký, lưu ký, bù trừ và thanh
toán giao dịch chứng khoán. - Đối với chứng khoán của công ty đại chúng đã
đăng ký tập trung tại VSD nhưng chưa, không niêm yết, đăng ký giao dịch
tại các SGDCK, VSD thực hiện chuyển quyền sở hữu theo văn bản hướng dẫn của UBCKNN.
Trên đây là nội dung quy định về nguyên tắc chuyển quyền sở hữu chứng khoán. Để
hiểu rõ hơn về vấn đề này, bạn nên tham khảo thêm tại Quyết định 196/QĐ-VSD năm
2017. Trân trọng!'
- 'Tại Khoản 2 Điều 27 Quy trình ban hành kèm theo Quyết định 595/QĐ-BHXH năm 2017,
có quy định: Cấp lại sổ BHXH do thay đổi họ, tên, chữ đệm; ngày, tháng, năm sinh;
giới tính, dân tộc; quốc tịch; điều chỉnh nội dung đã ghi trên sổ BHXH => Như
vậy, theo quy định nêu trên thì khi thay đổi họ, tên, chữ đệm trên sổ BHXH thì
sẽ làm thủ tục cấp lại. Chứ không làm thủ tục điều chỉnh trên sổ BHXH. Hồ sơ cấp
lại được quy định như sau: Thành phần hồ sơ Trường hợp người tham gia BHXH tự
làm - Tờ khai tham gia, điều chỉnh thông tin BHXH, BHYT (Mẫu TK1-TS) . - Hồ sơ
kèm theo (Mục 3, 4 Phụ lục 01) . Trường hợp đơn vị sử dụng làm : Bảng kê thông
tin (Mẫu D01-TS) . Số lượng hồ sơ: 01 bộ. Thời hạn giải quyết hồ sơ: Cấp lại sổ
BHXH do thay đổi họ, tên, chữ đệm; ngày, tháng, năm sinh; giới tính, dân tộc;
quốc tịch; sổ BHXH do mất, hỏng; cộng nối thời gian nhưng không phải đóng BHXH
hoặc gộp sổ BHXH: không quá 10 ngày kể từ ngày nhận đủ hồ sơ theo quy định. Trường
hợp cần phải xác minh quá trình đóng BHXH ở tỉnh khác hoặc nhiều đơn vị nơi người
lao động có thời gian làm việc thì không quá 45 ngày nhưng phải có văn bản thông
báo cho người lao động biết. Trân trọng.'
- 'Sổ bảo hiểm xã hội được cấp lại trong những trường hợp quy định tại Khoản 2 Điều
46 Quyết định 595/QĐ-BHXH năm 2017 Quy trình thu bảo hiểm xã hội, bảo hiểm y tế,
bảo hiểm thất nghiệp, bảo hiểm tai nạn lao động, bệnh nghề nghiệp; cấp sổ bảo
hiểm xã hội, thẻ bảo hiểm y tế do Bảo hiểm xã hội Việt Nam ban hành, bao gồm:
- Cấp lại sổ BHXH (bìa và tờ rời) các trường hợp: mất, hỏng; gộp; thay đổi số
sổ; họ, tên, chữ đệm; ngày, tháng, năm sinh; người đã hưởng BHXH một lần còn thời
gian đóng BHTN chưa hưởng. - Cấp lại bìa sổ BHXH các trường hợp: sai giới tính,
quốc tịch. - Cấp lại tờ rời sổ BHXH các trường hợp: mất, hỏng. Ban biên tập phản
hồi thông tin đến bạn.'
datasets:
- batmangiaicuuthegioi/augmentated_legal_triplets
pipeline_tag: sentence-similarity
library_name: sentence-transformers
---
# SentenceTransformer based on BAAI/bge-m3
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [BAAI/bge-m3](https://huggingface.co/BAAI/bge-m3) on the [augmentated_legal_triplets](https://huggingface.co/datasets/batmangiaicuuthegioi/augmentated_legal_triplets) dataset. It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
## Model Details
### Model Description
- **Model Type:** Sentence Transformer
- **Base model:** [BAAI/bge-m3](https://huggingface.co/BAAI/bge-m3) <!-- at revision 5617a9f61b028005a4858fdac845db406aefb181 -->
- **Maximum Sequence Length:** 8192 tokens
- **Output Dimensionality:** 1024 dimensions
- **Similarity Function:** Cosine Similarity
- **Training Dataset:**
- [augmentated_legal_triplets](https://huggingface.co/datasets/batmangiaicuuthegioi/augmentated_legal_triplets)
<!-- - **Language:** Unknown -->
<!-- - **License:** Unknown -->
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 8192, 'do_lower_case': False}) with Transformer model: XLMRobertaModel
(1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): Normalize()
)
```
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("batmangiaicuuthegioi/NODUPLICATES_512_10000")
# Run inference
sentences = [
'Chào Ban biên tập, tôi là Minh Tùng, hiện tôi đang làm việc tại một cơ quan hành chính nhà nước. Tôi có thắc mắc sau mong sớm nhận phản hồi. Cụ thể: Sổ bảo hiểm xã hội được cấp lại trong những trường hợp nào?',
'Sổ bảo hiểm xã hội được cấp lại trong những trường hợp quy định tại Khoản 2 Điều 46 Quyết định 595/QĐ-BHXH năm 2017 Quy trình thu bảo hiểm xã hội, bảo hiểm y tế, bảo hiểm thất nghiệp, bảo hiểm tai nạn lao động, bệnh nghề nghiệp; cấp sổ bảo hiểm xã hội, thẻ bảo hiểm y tế do Bảo hiểm xã hội Việt Nam ban hành, bao gồm: - Cấp lại sổ BHXH (bìa và tờ rời) các trường hợp: mất, hỏng; gộp; thay đổi số sổ; họ, tên, chữ đệm; ngày, tháng, năm sinh; người đã hưởng BHXH một lần còn thời gian đóng BHTN chưa hưởng. - Cấp lại bìa sổ BHXH các trường hợp: sai giới tính, quốc tịch. - Cấp lại tờ rời sổ BHXH các trường hợp: mất, hỏng. Ban biên tập phản hồi thông tin đến bạn.',
'Tại Khoản 2 Điều 27 Quy trình ban hành kèm theo Quyết định 595/QĐ-BHXH năm 2017, có quy định: Cấp lại sổ BHXH do thay đổi họ, tên, chữ đệm; ngày, tháng, năm sinh; giới tính, dân tộc; quốc tịch; điều chỉnh nội dung đã ghi trên sổ BHXH => Như vậy, theo quy định nêu trên thì khi thay đổi họ, tên, chữ đệm trên sổ BHXH thì sẽ làm thủ tục cấp lại. Chứ không làm thủ tục điều chỉnh trên sổ BHXH. Hồ sơ cấp lại được quy định như sau: Thành phần hồ sơ Trường hợp người tham gia BHXH tự làm - Tờ khai tham gia, điều chỉnh thông tin BHXH, BHYT (Mẫu TK1-TS) . - Hồ sơ kèm theo (Mục 3, 4 Phụ lục 01) . Trường hợp đơn vị sử dụng làm : Bảng kê thông tin (Mẫu D01-TS) . Số lượng hồ sơ: 01 bộ. Thời hạn giải quyết hồ sơ: Cấp lại sổ BHXH do thay đổi họ, tên, chữ đệm; ngày, tháng, năm sinh; giới tính, dân tộc; quốc tịch; sổ BHXH do mất, hỏng; cộng nối thời gian nhưng không phải đóng BHXH hoặc gộp sổ BHXH: không quá 10 ngày kể từ ngày nhận đủ hồ sơ theo quy định. Trường hợp cần phải xác minh quá trình đóng BHXH ở tỉnh khác hoặc nhiều đơn vị nơi người lao động có thời gian làm việc thì không quá 45 ngày nhưng phải có văn bản thông báo cho người lao động biết. Trân trọng.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
```
<!--
### Direct Usage (Transformers)
<details><summary>Click to see the direct usage in Transformers</summary>
</details>
-->
<!--
### Downstream Usage (Sentence Transformers)
You can finetune this model on your own dataset.
<details><summary>Click to expand</summary>
</details>
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Dataset
#### augmentated_legal_triplets
* Dataset: [augmentated_legal_triplets](https://huggingface.co/datasets/batmangiaicuuthegioi/augmentated_legal_triplets) at [e4cf34f](https://huggingface.co/datasets/batmangiaicuuthegioi/augmentated_legal_triplets/tree/e4cf34fadd27544b8e630479bd29ffc5fb2fdaa1)
* Size: 36,728 training samples
* Columns: <code>anchor</code>, <code>positive</code>, and <code>negative</code>
* Approximate statistics based on the first 1000 samples:
| | anchor | positive | negative |
|:--------|:-----------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|
| type | string | string | string |
| details | <ul><li>min: 8 tokens</li><li>mean: 66.04 tokens</li><li>max: 891 tokens</li></ul> | <ul><li>min: 32 tokens</li><li>mean: 314.94 tokens</li><li>max: 1194 tokens</li></ul> | <ul><li>min: 40 tokens</li><li>mean: 346.6 tokens</li><li>max: 751 tokens</li></ul> |
* Samples:
| anchor | positive | negative |
|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| <code>Luật Đất đai 2024 quy định thời hạn sử dụng đất xây dựng trụ sở cơ quan là bao lâu? Đất xây dựng trụ sở cơ quan thuộc nhóm đất nào?</code> | <code>Luật Đất đai 2024 quy định thời hạn sử dụng đất xây dựng trụ sở cơ quan là bao lâu? Căn cứ theo khoản 5 Điều 171 Luật Đất đai 2024 quy định như sau: Điều 171. Đất sử dụng ổn định lâu dài 1. Đất ở. 2. Đất nông nghiệp do cộng đồng dân cư sử dụng quy định tại khoản 4 Điều 178 của Luật này. 3. Đất rừng đặc dụng; đất rừng phòng hộ; đất rừng sản xuất do tổ chức quản lý. 4. Đất thương mại, dịch vụ, đất cơ sở sản xuất phi nông nghiệp của cá nhân đang sử dụng ổn định được Nhà nước công nhận mà không phải là đất được Nhà nước giao có thời hạn, cho thuê. 5. Đất xây dựng trụ sở cơ quan quy định tại khoản 1 Điều 199 của Luật này; đất xây dựng công trình sự nghiệp của đơn vị sự nghiệp công lập quy định tại khoản 2 Điều 199 của Luật này. 6. Đất quốc phòng, an ninh. 7. Đất tín ngưỡng. 8. Đất tôn giáo quy định tại khoản 2 Điều 213 của Luật này. 9. Đất sử dụng vào mục đích công cộng không có mục đích kinh doanh. 10. Đất nghĩa trang, nhà tang lễ, cơ sở hoả táng; đất cơ sở lưu giữ tro cốt. 11. Đất quy địn...</code> | <code>Đất ở tại đô thị có ký hiệu là gì? Theo khoản 2 Điều 10 Luật Đất đai 2013 quy định về việc phân loại đất như sau: Phân loại đất . .. 2. Nhóm đất phi nông nghiệp bao gồm các loại đất sau đây: a) Đất ở gồm đất ở tại nông thôn, đất ở tại đô thị; b) Đất xây dựng trụ sở cơ quan; c) Đất sử dụng vào mục đích quốc phòng, an ninh; d) Đất xây dựng công trình sự nghiệp gồm đất xây dựng trụ sở của tổ chức sự nghiệp; đất xây dựng cơ sở văn hoá, xã hội, y tế, giáo dục và đào tạo, thể dục thể thao, khoa học và công nghệ, ngoại giao và công trình sự nghiệp khác; đ) Đất sản xuất, kinh doanh phi nông nghiệp gồm đất khu công nghiệp, cụm công nghiệp, khu chế xuất; đất thương mại, dịch vụ; đất cơ sở sản xuất phi nông nghiệp; đất sử dụng cho hoạt động khoáng sản; đất sản xuất vật liệu xây dựng, làm đồ gốm; . .. Trong đó, theo Điều 144 Luật Đất đai 2013 quy định đất ở đô thị bao gồm: Đất ở tại đô thị 1. Đất ở tại đô thị bao gồm đất để xây dựng nhà ở, xây dựng các công trình phục vụ đời sống, vườn, ao trong c...</code> |
| <code>Ba tôi mất năm 2015 có để lại di chúc cho tôi 1 căn nhà. Ba tôi có 4 người con, mẹ tôi thì bệnh tâm thần. Cho tôi hỏi, để thực hiện được di chúc đó thì tôi cần làm thủ tục như thế nào?</code> | <code>Trường hợp thứ nhất, di chúc trên có hiệu lực pháp luật thì di sản thừa kế sẽ được phân chia theo di chúc, bạn có thể căn cứ theo quy định từ điều 652 đến 661 Bộ luật dân sự 2005 và loại di chúc của cha bạn để xác định di trên có hiệu lực pháp luật hay không. Điều 669 Bộ luật dân sự 2005 quy định về những người thừa kế không phụ thuộc vào nội dung của di chúc gồm: “Những người sau đây vẫn được hưởng phần di sản bằng hai phần ba suất của một người thừa kế theo pháp luật, nếu di sản được chia theo pháp luật, trong trường hợp họ không được người lập di chúc cho hưởng di sản hoặc chỉ cho hưởng phần di sản ít hơn hai phần ba suất đó, trừ khi họ là những người từ chối nhận di sản theo quy định tại Điều 642 hoặc họ là những người không có quyền hưởng di sản theo quy định tại khoản 1 Điều 643 của Bộ luật này: 1. Con chưa thành niên, cha, mẹ, vợ, chồng; 2. Con đã thành niên mà không có khả năng lao động. ” Như vậy, mẹ bạn thuộc trường hợp người thừa kế không phụ thuộc vào nội dung của di chúc n...</code> | <code>Theo khoản 1 a Điều 675, 1a Điều 676 Bộ luật Dân sự, khi cha bạn mất không có di chúc thì mẹ và các anh em của bạn cùng được thừa kế theo pháp luật phần di sản của người chết theo nguyên tắc mỗi người được hưởng phần di sản bằng nhau. Nếu không thể thoả thuận với một đồng thừa kế về việc bán nhà thuộc sở hữu chung, ba thành viên còn lại có thể khởi kiện đến TAND cấp huyện nơi có căn nhà để được xem xét, giải quyết. Trường hợp muốn được sở hữu riêng căn nhà, một thành viên có trách nhiệm thanh toán cho những người đồng sở hữu khác phần trị giá nhà mà họ được hưởng.</code> |
| <code>Kính gửi quý Luật sư, Tôi cùng bạn tôi có dự định thành lập công ty TNHH 1 thành viên (là cá nhân) chuyên thực hiện các nghiệp vụ XNK (nhập và xuất hàng). Như vậy, xin quý luật sư tư vấn giúp khi thực hiện đăng ký giấy phép kinh doanh chúng tôi sẽ đăng ký hình thức kinh doanh là gì? Vì tôi có nghiên cứu qua các bài tư vấn trước rằng sau khi thành lập công ty để thực hiện nghiệp vụ XNK còn cần đến việc đăng ký giấy phép XNK. Trong trường hợp công ty trong dự kiến của chúng tôi chỉ chuyên về việc thực hiện các nghiệp vụ XNK ủy thác từ các công ty khác thì nên đăng ký loại hình nào là phù hợp? Vì chúng tôi có sẵn đối tác ổn định là công ty đặt tại nước ngoài và mặt hàng dự định xuất không giới hạn (đa số là linh kiện điện tử do VN gia công, sản xuất). Chúng tôi có quyền thu mua và xuất hàng hay không nếu là công ty XNK?</code> | <code>Chào Bạn, 1. Nếu Bạn và Bạn của bạn (người Việt Nam) thành lập Công ty TNHH thì phải là 2 thành viên chứ không phải 1 thành viên ! Theo như bạn trình bày thì bạn có thể đăng ký ngành nghề là dịch vụ Logistic (Quy định tại Điều 233 Luật Thương mại) Bạn có thể tra cứu tại QĐ 10/2007/QĐ-TTg hoặc tìm hiểu thêm với Cán bộ ở Sở Kế hoạch Đầu tư để biết mã ngành của dịch vụ này. 2. Về quyền xuất nhập khẩu của thương nhân Việt Nam. Nghị định 12/2006/NĐ-CP quy định như sau: Điều 3. Quyền kinh doanh xuất khẩu, nhập khẩu 1. Đối với thương nhân Việt Nam không có vốn đầu tư trực tiếp của nước ngoài (dưới đây gọi tắt là thương nhân) : Trừ hàng hoá thuộc Danh mục cấm xuất khẩu, tạm ngừng xuất khẩu, hàng hoá thuộc Danh mục cấm nhập khẩu, tạm ngừng nhập khẩu, thương nhân được xuất khẩu nhập khẩu hàng hoá không phụ thuộc vào ngành nghề đăng ký kinh doanh. 3. Nếu muốn thực hiện chức năng mua bán linh kiện điện tử thì bạn đăng ký với Sở kế hoạch Đầu tư. Chúc may mắn.</code> | <code>Điều kiện cho người nước ngoài kinh doanh hoạt động xuất nhập khẩu ở Việt Nam? Căn cứ Khoản 1, Khoản 2 Điều 17 Luật Doanh nghiệp 2020 về quyền thành lập, góp vốn, mua cổ phần, mua phần vốn góp và quản lý doanh nghiệp như sau: 1. Tổ chức, cá nhân có quyền thành lập và quản lý doanh nghiệp tại Việt Nam theo quy định của Luật này, trừ trường hợp quy định tại khoản 2 Điều này. 2. Tổ chức, cá nhân sau đây không có quyền thành lập và quản lý doanh nghiệp tại Việt Nam: a) Cơ quan nhà nước, đơn vị lực lượng vũ trang nhân dân sử dụng tài sản nhà nước để thành lập doanh nghiệp kinh doanh thu lợi riêng cho cơ quan, đơn vị mình; b) Cán bộ, công chức, viên chức theo quy định của Luật Cán bộ, công chức và Luật Viên chức; c) Sĩ quan, hạ sĩ quan, quân nhân chuyên nghiệp, công nhân, viên chức quốc phòng trong các cơ quan, đơn vị thuộc Quân đội nhân dân Việt Nam; sĩ quan, hạ sĩ quan chuyên nghiệp, công nhân công an trong các cơ quan, đơn vị thuộc Công an nhân dân Việt Nam, trừ người được cử làm đại diện...</code> |
* Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
```json
{
"scale": 20.0,
"similarity_fct": "cos_sim"
}
```
### Evaluation Dataset
#### augmentated_legal_triplets
* Dataset: [augmentated_legal_triplets](https://huggingface.co/datasets/batmangiaicuuthegioi/augmentated_legal_triplets) at [e4cf34f](https://huggingface.co/datasets/batmangiaicuuthegioi/augmentated_legal_triplets/tree/e4cf34fadd27544b8e630479bd29ffc5fb2fdaa1)
* Size: 459 evaluation samples
* Columns: <code>anchor</code>, <code>positive</code>, and <code>negative</code>
* Approximate statistics based on the first 459 samples:
| | anchor | positive | negative |
|:--------|:-----------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|
| type | string | string | string |
| details | <ul><li>min: 8 tokens</li><li>mean: 67.54 tokens</li><li>max: 596 tokens</li></ul> | <ul><li>min: 30 tokens</li><li>mean: 317.5 tokens</li><li>max: 696 tokens</li></ul> | <ul><li>min: 2 tokens</li><li>mean: 353.73 tokens</li><li>max: 715 tokens</li></ul> |
* Samples:
| anchor | positive | negative |
|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| <code>Luật Thư viện 2019 quy định Thư viện tư nhân có phục vụ cộng đồng là gì?</code> | <code>Điều 16. Thư viện cộng đồng và thư viện tư nhân có phục vụ cộng đồng. khoản 1. thư viện cộng đồng là thư viện, có tài nguyên thông tin tổng hợp do cộng đồng dân cư thành lập tại trung tâm học tập cộng đồng, trung tâm văn hóa, thể thao xã, phường, thị trấn; điểm bưu điện văn hóa xã; nhà văn hóa thôn, ấp, bản, làng, buôn, phum, sóc; khu chung cư; nơi sinh hoạt chung của cộng đồng. Thư viện cộng đồng và thư viện tư nhân có phục vụ cộng đồng. khoản 2. thư viện tư nhân có phục vụ cộng đồng là thư viện có tài nguyên thông tin tổng hợp hoặc chuyên ngành do tổ chức, cá nhân việt nam thành lập, tự bảo đảm kinh phí hoạt động. Thư viện cộng đồng và thư viện tư nhân có phục vụ cộng đồng.</code> | <code>Điều 2. Đối tượng áp dụng. điểm b) ủy ban nhân dân các cấp; các cơ quan chuyên môn thuộc ủy ban nhân dân cấp tỉnh, cấp huyện; các ban quản lý khu công nghiệp, khu chế xuất, khu kinh tế, khu công nghệ cao có nhiệm vụ, quyền hạn giải quyết thủ tục hành chính. các cơ quan quy định tại điểm a, b khoản này sau đây được gọi tắt là cơ quan có thẩm quyền. Đối tượng áp dụng. khoản 2. cán bộ, công chức, viên chức của các cơ quan có thẩm quyền quy định tại khoản 1 điều này, sỹ quan, hạ sỹ quan quân đội, sỹ quan, hạ sỹ quan công an thuộc bộ quốc phòng, bộ công an (sau đây gọi chung là cán bộ, công chức, viên chức). Đối tượng áp dụng. khoản 3. tổ chức, cá nhân có yêu cầu thực</code> |
| <code>Thế nào là phương thức khớp lệnh tập trung?</code> | <code>Điều 2. Giải thích từ ngữ. trong thông tư này, các từ ngữ dưới đây được hiểu như sau: Giải thích từ ngữ. khoản 1. biên độ dao động giá là giới hạn dao động giá chứng khoán quy định trong ngày giao dịch được tính theo tỷ lệ phần trăm (%) so với giá tham chiếu. Giải thích từ ngữ. khoản 2. ngắt mạch thị trường (circuit breaker) là cơ chế tạm dừng giao dịch tự động trong phiên giao dịch khi giá chứng khoán hoặc chỉ số chứng khoán biến động chạm các ngưỡng xác định trên hệ thống giao dịch chứng khoán. Giải thích từ ngữ. khoản 3. giá tham chiếu là mức giá do sở giao dịch chứng khoán xác định và được dùng làm cơ sở để xác định giá cao nhất (giá trần), giá thấp nhất (giá sàn) trong ngày giao dịch. Giải thích</code> | <code>Điều 42. Trình tự, thủ tục xử lý kỷ luật trong một số trường hợp đặc biệt. khoản 1. trường hợp người vi phạm thuộc quyền có hành vi chống mệnh lệnh hoặc có hành vi vi phạm pháp luật nghiêm trọng thì người chỉ huy phải có biện pháp ngăn chặn kịp thời và báo cáo ngay lên cấp trên có thẩm quyền. Trình tự, thủ tục xử lý kỷ luật trong một số trường hợp đặc biệt. khoản 2. trường hợp người vi phạm không chấp hành kiểm điểm xét kỷ luật thì người chỉ huy căn cứ tính chất, mức độ của hành vi vi phạm, đề nghị của cấp dưới và các tổ chức quần chúng để triệu tập họp chỉ huy, cấp ủy xem xét, quyết định hình thức kỷ luật theo quyền hạn. Trình tự, thủ tục xử lý kỷ luật trong một số</code> |
| <code>Đơn giản hóa thủ tục khai thuế TNCN đối với cá nhân nhận thừa kế, quà tặng là nhà ở, công trình xây dựng hình thành trong tương lai tại Việt Nam được quy định như thế nào? Xin chào Ban biên tập, tôi là Bảo Long hiện đang sống và làm việc Bến Tre. Tôi đang tìm hiểu về thủ tục hành chính trong lĩnh vực thuế. Tôi có nghe nói về việc đơn giản hóa thủ tục hành chính trong lĩnh vực thuế. Vậy Ban biên tập cho tôi hỏi đơn giản hóa thủ tục khai thuế TNCN đối với cá nhân nhận thừa kế, quà tặng là nhà ở, công trình xây dựng hình thành trong tương lai tại Việt Nam được quy định như thế nào? Vấn đề này được quy định cụ thể tại văn bản nào?</code> | <code>Đơn giản hoá thủ tục khai thuế TNCN đối với cá nhân nhận thừa kế, quà tặng là nhà ở, công trình xây dựng hình thành trong tương lai tại Việt Nam được quy định tại Tiểu mục 21 Mục III Phần A Phương án đơn giản hoá thủ tục hành chính, giấy tờ công dân liên quan đến quản lý dân cư thuộc phạm vi chức năng quản lý của Bộ Tài chính ban hành kèm theo Nghị quyết 104/NQ-CP năm 2017, cụ thể: Bỏ các chỉ tiêu từ [10] đến [12]: Địa chỉ, Quận/huyện, Tỉnh/thành phố (đối với cá nhân) và bổ sung thêm trường thông tin về số định danh cá nhân tại Tờ khai mẫu số 03/BĐS-TNCN ban hành kèm theo Thông tư số 92/2015/TT-BTC. Trên đây là tư vấn về đơn giản hoá thủ tục khai thuế TNCN đối với cá nhân nhận thừa kế, quà tặng là nhà ở, công trình xây dựng hình thành trong tương lai tại Việt Nam. Để biết thêm thông tin chi tiết bạn nên tham khảo tại Nghị quyết 104/NQ-CP năm 2017. Mong rằng những tư vấn của chúng tôi sẽ giúp giải đáp được những vướng mắc của bạn. Chào thân ái và chúc sức khoẻ!</code> | <code>Căn cứ tính thuế thu nhập cá nhân đối với thu nhập từ thừa kế, quà tặng được quy định tại Điều 16 Thông tư 111/2013/TT-BTC Hướng dẫn Luật thuế thu nhập cá nhân và Nghị định 65/2013/NĐ-CP do Bộ trưởng Bộ Tài chính ban hành như sau: Căn cứ tính thuế đối với thu nhập từ thừa kế, quà tặng là thu nhập tính thuế và thuế suất. 1. Thu nhập tính thuế Thu nhập tính thuế từ nhận thừa kế, quà tặng là phần giá trị tài sản nhận thừa kế, quà tặng vượt trên 10 triệu đồng mỗi lần nhận. Giá trị tài sản nhận thừa kế, quà tặng được xác định đối với từng trường hợp, cụ thể như sau: a) Đối với thừa kế, quà tặng là chứng khoán: giá trị tài sản nhận thừa kế là giá trị chứng khoán tại thời điểm đăng ký chuyển quyền sở hữu. Thu nhập tính thuế từ thừa kế, quà tặng là chứng khoán là phần giá trị tài sản nhận thừa kế, quà tặng vượt trên 10 triệu đồng tính trên toàn bộ các mã chứng khoán nhận được chưa trừ bất cứ một khoản chi phí nào. Cụ thể như sau: (Điểm này được sửa đổi bởi Khoản 1 Điều 19 Thông tư 92/2015/TT-B...</code> |
* Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
```json
{
"scale": 20.0,
"similarity_fct": "cos_sim"
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `eval_strategy`: steps
- `per_device_train_batch_size`: 2
- `per_device_eval_batch_size`: 1
- `learning_rate`: 5e-06
- `num_train_epochs`: 1
- `max_steps`: 10000
- `warmup_ratio`: 0.1
- `fp16`: True
- `batch_sampler`: no_duplicates
#### All Hyperparameters
<details><summary>Click to expand</summary>
- `overwrite_output_dir`: False
- `do_predict`: False
- `eval_strategy`: steps
- `prediction_loss_only`: True
- `per_device_train_batch_size`: 2
- `per_device_eval_batch_size`: 1
- `per_gpu_train_batch_size`: None
- `per_gpu_eval_batch_size`: None
- `gradient_accumulation_steps`: 1
- `eval_accumulation_steps`: None
- `torch_empty_cache_steps`: None
- `learning_rate`: 5e-06
- `weight_decay`: 0.0
- `adam_beta1`: 0.9
- `adam_beta2`: 0.999
- `adam_epsilon`: 1e-08
- `max_grad_norm`: 1.0
- `num_train_epochs`: 1
- `max_steps`: 10000
- `lr_scheduler_type`: linear
- `lr_scheduler_kwargs`: {}
- `warmup_ratio`: 0.1
- `warmup_steps`: 0
- `log_level`: passive
- `log_level_replica`: warning
- `log_on_each_node`: True
- `logging_nan_inf_filter`: True
- `save_safetensors`: True
- `save_on_each_node`: False
- `save_only_model`: False
- `restore_callback_states_from_checkpoint`: False
- `no_cuda`: False
- `use_cpu`: False
- `use_mps_device`: False
- `seed`: 42
- `data_seed`: None
- `jit_mode_eval`: False
- `use_ipex`: False
- `bf16`: False
- `fp16`: True
- `fp16_opt_level`: O1
- `half_precision_backend`: auto
- `bf16_full_eval`: False
- `fp16_full_eval`: False
- `tf32`: None
- `local_rank`: 0
- `ddp_backend`: None
- `tpu_num_cores`: None
- `tpu_metrics_debug`: False
- `debug`: []
- `dataloader_drop_last`: False
- `dataloader_num_workers`: 0
- `dataloader_prefetch_factor`: None
- `past_index`: -1
- `disable_tqdm`: False
- `remove_unused_columns`: True
- `label_names`: None
- `load_best_model_at_end`: False
- `ignore_data_skip`: False
- `fsdp`: []
- `fsdp_min_num_params`: 0
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
- `fsdp_transformer_layer_cls_to_wrap`: None
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
- `deepspeed`: None
- `label_smoothing_factor`: 0.0
- `optim`: adamw_torch
- `optim_args`: None
- `adafactor`: False
- `group_by_length`: False
- `length_column_name`: length
- `ddp_find_unused_parameters`: None
- `ddp_bucket_cap_mb`: None
- `ddp_broadcast_buffers`: False
- `dataloader_pin_memory`: True
- `dataloader_persistent_workers`: False
- `skip_memory_metrics`: True
- `use_legacy_prediction_loop`: False
- `push_to_hub`: False
- `resume_from_checkpoint`: None
- `hub_model_id`: None
- `hub_strategy`: every_save
- `hub_private_repo`: None
- `hub_always_push`: False
- `gradient_checkpointing`: False
- `gradient_checkpointing_kwargs`: None
- `include_inputs_for_metrics`: False
- `include_for_metrics`: []
- `eval_do_concat_batches`: True
- `fp16_backend`: auto
- `push_to_hub_model_id`: None
- `push_to_hub_organization`: None
- `mp_parameters`:
- `auto_find_batch_size`: False
- `full_determinism`: False
- `torchdynamo`: None
- `ray_scope`: last
- `ddp_timeout`: 1800
- `torch_compile`: False
- `torch_compile_backend`: None
- `torch_compile_mode`: None
- `include_tokens_per_second`: False
- `include_num_input_tokens_seen`: False
- `neftune_noise_alpha`: None
- `optim_target_modules`: None
- `batch_eval_metrics`: False
- `eval_on_start`: False
- `use_liger_kernel`: False
- `eval_use_gather_object`: False
- `average_tokens_across_devices`: False
- `prompts`: None
- `batch_sampler`: no_duplicates
- `multi_dataset_batch_sampler`: proportional
</details>
### Training Logs
<details><summary>Click to expand</summary>
| Epoch | Step | Training Loss | Validation Loss |
|:------:|:-----:|:-------------:|:---------------:|
| 0.0005 | 10 | 0.1714 | - |
| 0.0011 | 20 | 0.1622 | - |
| 0.0016 | 30 | 0.2432 | - |
| 0.0022 | 40 | 0.1748 | - |
| 0.0027 | 50 | 0.1608 | - |
| 0.0033 | 60 | 0.2183 | - |
| 0.0038 | 70 | 0.1434 | - |
| 0.0044 | 80 | 0.2653 | - |
| 0.0049 | 90 | 0.1565 | - |
| 0.0054 | 100 | 0.1643 | - |
| 0.0060 | 110 | 0.1897 | - |
| 0.0065 | 120 | 0.2513 | - |
| 0.0071 | 130 | 0.1992 | - |
| 0.0076 | 140 | 0.2166 | - |
| 0.0082 | 150 | 0.2571 | - |
| 0.0087 | 160 | 0.1926 | - |
| 0.0093 | 170 | 0.2111 | - |
| 0.0098 | 180 | 0.1246 | - |
| 0.0103 | 190 | 0.239 | - |
| 0.0109 | 200 | 0.1129 | - |
| 0.0114 | 210 | 0.1357 | - |
| 0.0120 | 220 | 0.3692 | - |
| 0.0125 | 230 | 0.1737 | - |
| 0.0131 | 240 | 0.1115 | - |
| 0.0136 | 250 | 0.2765 | - |
| 0.0142 | 260 | 0.1278 | - |
| 0.0147 | 270 | 0.2341 | - |
| 0.0152 | 280 | 0.1628 | - |
| 0.0158 | 290 | 0.0902 | - |
| 0.0163 | 300 | 0.1395 | - |
| 0.0169 | 310 | 0.188 | - |
| 0.0174 | 320 | 0.1639 | - |
| 0.0180 | 330 | 0.0677 | - |
| 0.0185 | 340 | 0.1283 | - |
| 0.0191 | 350 | 0.1618 | - |
| 0.0196 | 360 | 0.101 | - |
| 0.0201 | 370 | 0.1689 | - |
| 0.0207 | 380 | 0.4551 | - |
| 0.0212 | 390 | 0.1724 | - |
| 0.0218 | 400 | 0.0459 | - |
| 0.0223 | 410 | 0.1484 | - |
| 0.0229 | 420 | 0.048 | - |
| 0.0234 | 430 | 0.1644 | - |
| 0.0240 | 440 | 0.0682 | - |
| 0.0245 | 450 | 0.195 | - |
| 0.0250 | 460 | 0.1838 | - |
| 0.0256 | 470 | 0.0961 | - |
| 0.0261 | 480 | 0.0971 | - |
| 0.0267 | 490 | 0.0612 | - |
| 0.0272 | 500 | 0.0326 | 0.0597 |
| 0.0278 | 510 | 0.2155 | - |
| 0.0283 | 520 | 0.1117 | - |
| 0.0289 | 530 | 0.2924 | - |
| 0.0294 | 540 | 0.0415 | - |
| 0.0299 | 550 | 0.1668 | - |
| 0.0305 | 560 | 0.0458 | - |
| 0.0310 | 570 | 0.1168 | - |
| 0.0316 | 580 | 0.3376 | - |
| 0.0321 | 590 | 0.0247 | - |
| 0.0327 | 600 | 0.1134 | - |
| 0.0332 | 610 | 0.055 | - |
| 0.0338 | 620 | 0.0211 | - |
| 0.0343 | 630 | 0.0567 | - |
| 0.0349 | 640 | 0.0303 | - |
| 0.0354 | 650 | 0.0449 | - |
| 0.0359 | 660 | 0.105 | - |
| 0.0365 | 670 | 0.0749 | - |
| 0.0370 | 680 | 0.0256 | - |
| 0.0376 | 690 | 0.0996 | - |
| 0.0381 | 700 | 0.0687 | - |
| 0.0387 | 710 | 0.1861 | - |
| 0.0392 | 720 | 0.0363 | - |
| 0.0398 | 730 | 0.305 | - |
| 0.0403 | 740 | 0.0703 | - |
| 0.0408 | 750 | 0.0568 | - |
| 0.0414 | 760 | 0.0203 | - |
| 0.0419 | 770 | 0.0348 | - |
| 0.0425 | 780 | 0.1217 | - |
| 0.0430 | 790 | 0.0644 | - |
| 0.0436 | 800 | 0.4476 | - |
| 0.0441 | 810 | 0.0679 | - |
| 0.0447 | 820 | 0.0357 | - |
| 0.0452 | 830 | 0.3741 | - |
| 0.0457 | 840 | 0.1271 | - |
| 0.0463 | 850 | 0.0998 | - |
| 0.0468 | 860 | 0.0452 | - |
| 0.0474 | 870 | 0.0157 | - |
| 0.0479 | 880 | 0.0125 | - |
| 0.0485 | 890 | 0.1142 | - |
| 0.0490 | 900 | 0.3188 | - |
| 0.0496 | 910 | 0.172 | - |
| 0.0501 | 920 | 0.0973 | - |
| 0.0506 | 930 | 0.2071 | - |
| 0.0512 | 940 | 0.0492 | - |
| 0.0517 | 950 | 0.0198 | - |
| 0.0523 | 960 | 0.0647 | - |
| 0.0528 | 970 | 0.0094 | - |
| 0.0534 | 980 | 0.0491 | - |
| 0.0539 | 990 | 0.0696 | - |
| 0.0545 | 1000 | 0.1467 | 0.0516 |
| 0.0550 | 1010 | 0.0433 | - |
| 0.0555 | 1020 | 0.0643 | - |
| 0.0561 | 1030 | 0.0377 | - |
| 0.0566 | 1040 | 0.2884 | - |
| 0.0572 | 1050 | 0.0253 | - |
| 0.0577 | 1060 | 0.0211 | - |
| 0.0583 | 1070 | 0.0413 | - |
| 0.0588 | 1080 | 0.025 | - |
| 0.0594 | 1090 | 0.0294 | - |
| 0.0599 | 1100 | 0.0123 | - |
| 0.0604 | 1110 | 0.1298 | - |
| 0.0610 | 1120 | 0.0924 | - |
| 0.0615 | 1130 | 0.0294 | - |
| 0.0621 | 1140 | 0.1025 | - |
| 0.0626 | 1150 | 0.0179 | - |
| 0.0632 | 1160 | 0.0534 | - |
| 0.0637 | 1170 | 0.061 | - |
| 0.0643 | 1180 | 0.0152 | - |
| 0.0648 | 1190 | 0.0675 | - |
| 0.0653 | 1200 | 0.027 | - |
| 0.0659 | 1210 | 0.0125 | - |
| 0.0664 | 1220 | 0.1197 | - |
| 0.0670 | 1230 | 0.1654 | - |
| 0.0675 | 1240 | 0.0626 | - |
| 0.0681 | 1250 | 0.0884 | - |
| 0.0686 | 1260 | 0.0386 | - |
| 0.0692 | 1270 | 0.0329 | - |
| 0.0697 | 1280 | 0.0831 | - |
| 0.0702 | 1290 | 0.003 | - |
| 0.0708 | 1300 | 0.0768 | - |
| 0.0713 | 1310 | 0.0737 | - |
| 0.0719 | 1320 | 0.1987 | - |
| 0.0724 | 1330 | 0.0888 | - |
| 0.0730 | 1340 | 0.0103 | - |
| 0.0735 | 1350 | 0.0184 | - |
| 0.0741 | 1360 | 0.0392 | - |
| 0.0746 | 1370 | 0.0392 | - |
| 0.0751 | 1380 | 0.2632 | - |
| 0.0757 | 1390 | 0.1563 | - |
| 0.0762 | 1400 | 0.1196 | - |
| 0.0768 | 1410 | 0.1273 | - |
| 0.0773 | 1420 | 0.0052 | - |
| 0.0779 | 1430 | 0.0497 | - |
| 0.0784 | 1440 | 0.0033 | - |
| 0.0790 | 1450 | 0.299 | - |
| 0.0795 | 1460 | 0.1286 | - |
| 0.0800 | 1470 | 0.1639 | - |
| 0.0806 | 1480 | 0.0263 | - |
| 0.0811 | 1490 | 0.0164 | - |
| 0.0817 | 1500 | 0.062 | 0.0506 |
| 0.0822 | 1510 | 0.0292 | - |
| 0.0828 | 1520 | 0.0011 | - |
| 0.0833 | 1530 | 0.0215 | - |
| 0.0839 | 1540 | 0.082 | - |
| 0.0844 | 1550 | 0.0037 | - |
| 0.0849 | 1560 | 0.0541 | - |
| 0.0855 | 1570 | 0.0301 | - |
| 0.0860 | 1580 | 0.0764 | - |
| 0.0866 | 1590 | 0.5301 | - |
| 0.0871 | 1600 | 0.0198 | - |
| 0.0877 | 1610 | 0.1147 | - |
| 0.0882 | 1620 | 0.0355 | - |
| 0.0888 | 1630 | 0.0205 | - |
| 0.0893 | 1640 | 0.0463 | - |
| 0.0898 | 1650 | 0.03 | - |
| 0.0904 | 1660 | 0.0905 | - |
| 0.0909 | 1670 | 0.1239 | - |
| 0.0915 | 1680 | 0.0059 | - |
| 0.0920 | 1690 | 0.0199 | - |
| 0.0926 | 1700 | 0.01 | - |
| 0.0931 | 1710 | 0.0093 | - |
| 0.0937 | 1720 | 0.0221 | - |
| 0.0942 | 1730 | 0.0908 | - |
| 0.0948 | 1740 | 0.1225 | - |
| 0.0953 | 1750 | 0.1216 | - |
| 0.0958 | 1760 | 0.0525 | - |
| 0.0964 | 1770 | 0.0384 | - |
| 0.0969 | 1780 | 0.0293 | - |
| 0.0975 | 1790 | 0.0454 | - |
| 0.0980 | 1800 | 0.0101 | - |
| 0.0986 | 1810 | 0.0098 | - |
| 0.0991 | 1820 | 0.0811 | - |
| 0.0997 | 1830 | 0.0628 | - |
| 0.1002 | 1840 | 0.0578 | - |
| 0.1007 | 1850 | 0.0339 | - |
| 0.1013 | 1860 | 0.0203 | - |
| 0.1018 | 1870 | 0.0216 | - |
| 0.1024 | 1880 | 0.0147 | - |
| 0.1029 | 1890 | 0.0168 | - |
| 0.1035 | 1900 | 0.0312 | - |
| 0.1040 | 1910 | 0.0689 | - |
| 0.1046 | 1920 | 0.0524 | - |
| 0.1051 | 1930 | 0.0158 | - |
| 0.1056 | 1940 | 0.0064 | - |
| 0.1062 | 1950 | 0.0705 | - |
| 0.1067 | 1960 | 0.0126 | - |
| 0.1073 | 1970 | 0.1654 | - |
| 0.1078 | 1980 | 0.0081 | - |
| 0.1084 | 1990 | 0.0182 | - |
| 0.1089 | 2000 | 0.0048 | 0.0489 |
| 0.1095 | 2010 | 0.0498 | - |
| 0.1100 | 2020 | 0.2631 | - |
| 0.1105 | 2030 | 0.0349 | - |
| 0.1111 | 2040 | 0.0325 | - |
| 0.1116 | 2050 | 0.0458 | - |
| 0.1122 | 2060 | 0.0297 | - |
| 0.1127 | 2070 | 0.037 | - |
| 0.1133 | 2080 | 0.0057 | - |
| 0.1138 | 2090 | 0.0469 | - |
| 0.1144 | 2100 | 0.0451 | - |
| 0.1149 | 2110 | 0.0024 | - |
| 0.1154 | 2120 | 0.3465 | - |
| 0.1160 | 2130 | 0.0191 | - |
| 0.1165 | 2140 | 0.0429 | - |
| 0.1171 | 2150 | 0.2277 | - |
| 0.1176 | 2160 | 0.0384 | - |
| 0.1182 | 2170 | 0.024 | - |
| 0.1187 | 2180 | 0.04 | - |
| 0.1193 | 2190 | 0.0127 | - |
| 0.1198 | 2200 | 0.082 | - |
| 0.1203 | 2210 | 0.237 | - |
| 0.1209 | 2220 | 0.0581 | - |
| 0.1214 | 2230 | 0.0185 | - |
| 0.1220 | 2240 | 0.0358 | - |
| 0.1225 | 2250 | 0.0635 | - |
| 0.1231 | 2260 | 0.0091 | - |
| 0.1236 | 2270 | 0.0484 | - |
| 0.1242 | 2280 | 0.0578 | - |
| 0.1247 | 2290 | 0.0503 | - |
| 0.1252 | 2300 | 0.0495 | - |
| 0.1258 | 2310 | 0.0591 | - |
| 0.1263 | 2320 | 0.0211 | - |
| 0.1269 | 2330 | 0.0091 | - |
| 0.1274 | 2340 | 0.0189 | - |
| 0.1280 | 2350 | 0.0173 | - |
| 0.1285 | 2360 | 0.0117 | - |
| 0.1291 | 2370 | 0.1363 | - |
| 0.1296 | 2380 | 0.0254 | - |
| 0.1301 | 2390 | 0.0351 | - |
| 0.1307 | 2400 | 0.0037 | - |
| 0.1312 | 2410 | 0.0414 | - |
| 0.1318 | 2420 | 0.0016 | - |
| 0.1323 | 2430 | 0.0116 | - |
| 0.1329 | 2440 | 0.0208 | - |
| 0.1334 | 2450 | 0.0134 | - |
| 0.1340 | 2460 | 0.0422 | - |
| 0.1345 | 2470 | 0.0082 | - |
| 0.1350 | 2480 | 0.1176 | - |
| 0.1356 | 2490 | 0.0293 | - |
| 0.1361 | 2500 | 0.0064 | 0.0427 |
| 0.1367 | 2510 | 0.0017 | - |
| 0.1372 | 2520 | 0.0736 | - |
| 0.1378 | 2530 | 0.3078 | - |
| 0.1383 | 2540 | 0.0799 | - |
| 0.1389 | 2550 | 0.0094 | - |
| 0.1394 | 2560 | 0.1236 | - |
| 0.1399 | 2570 | 0.0081 | - |
| 0.1405 | 2580 | 0.007 | - |
| 0.1410 | 2590 | 0.0191 | - |
| 0.1416 | 2600 | 0.011 | - |
| 0.1421 | 2610 | 0.2477 | - |
| 0.1427 | 2620 | 0.0809 | - |
| 0.1432 | 2630 | 0.0353 | - |
| 0.1438 | 2640 | 0.0108 | - |
| 0.1443 | 2650 | 0.0639 | - |
| 0.1448 | 2660 | 0.0023 | - |
| 0.1454 | 2670 | 0.0428 | - |
| 0.1459 | 2680 | 0.0305 | - |
| 0.1465 | 2690 | 0.0239 | - |
| 0.1470 | 2700 | 0.3158 | - |
| 0.1476 | 2710 | 0.0376 | - |
| 0.1481 | 2720 | 0.0158 | - |
| 0.1487 | 2730 | 0.0249 | - |
| 0.1492 | 2740 | 0.3198 | - |
| 0.1497 | 2750 | 0.0277 | - |
| 0.1503 | 2760 | 0.1732 | - |
| 0.1508 | 2770 | 0.0967 | - |
| 0.1514 | 2780 | 0.0564 | - |
| 0.1519 | 2790 | 0.0221 | - |
| 0.1525 | 2800 | 0.0412 | - |
| 0.1530 | 2810 | 0.0848 | - |
| 0.1536 | 2820 | 0.0194 | - |
| 0.1541 | 2830 | 0.0147 | - |
| 0.1547 | 2840 | 0.0088 | - |
| 0.1552 | 2850 | 0.0537 | - |
| 0.1557 | 2860 | 0.0142 | - |
| 0.1563 | 2870 | 0.009 | - |
| 0.1568 | 2880 | 0.2238 | - |
| 0.1574 | 2890 | 0.0081 | - |
| 0.1579 | 2900 | 0.0615 | - |
| 0.1585 | 2910 | 0.0225 | - |
| 0.1590 | 2920 | 0.2693 | - |
| 0.1596 | 2930 | 0.0245 | - |
| 0.1601 | 2940 | 0.0143 | - |
| 0.1606 | 2950 | 0.0027 | - |
| 0.1612 | 2960 | 0.0269 | - |
| 0.1617 | 2970 | 0.0277 | - |
| 0.1623 | 2980 | 0.0084 | - |
| 0.1628 | 2990 | 0.0034 | - |
| 0.1634 | 3000 | 0.4531 | 0.0394 |
| 0.1639 | 3010 | 0.0258 | - |
| 0.1645 | 3020 | 0.0465 | - |
| 0.1650 | 3030 | 0.047 | - |
| 0.1655 | 3040 | 0.0013 | - |
| 0.1661 | 3050 | 0.0369 | - |
| 0.1666 | 3060 | 0.0135 | - |
| 0.1672 | 3070 | 0.0398 | - |
| 0.1677 | 3080 | 0.0411 | - |
| 0.1683 | 3090 | 0.0147 | - |
| 0.1688 | 3100 | 0.0781 | - |
| 0.1694 | 3110 | 0.0074 | - |
| 0.1699 | 3120 | 0.0208 | - |
| 0.1704 | 3130 | 0.0433 | - |
| 0.1710 | 3140 | 0.0091 | - |
| 0.1715 | 3150 | 0.0133 | - |
| 0.1721 | 3160 | 0.067 | - |
| 0.1726 | 3170 | 0.1392 | - |
| 0.1732 | 3180 | 0.2033 | - |
| 0.1737 | 3190 | 0.0409 | - |
| 0.1743 | 3200 | 0.0197 | - |
| 0.1748 | 3210 | 0.0107 | - |
| 0.1753 | 3220 | 0.0084 | - |
| 0.1759 | 3230 | 0.0107 | - |
| 0.1764 | 3240 | 0.0618 | - |
| 0.1770 | 3250 | 0.0326 | - |
| 0.1775 | 3260 | 0.0033 | - |
| 0.1781 | 3270 | 0.1116 | - |
| 0.1786 | 3280 | 0.0644 | - |
| 0.1792 | 3290 | 0.0144 | - |
| 0.1797 | 3300 | 0.0838 | - |
| 0.1802 | 3310 | 0.0007 | - |
| 0.1808 | 3320 | 0.0373 | - |
| 0.1813 | 3330 | 0.0021 | - |
| 0.1819 | 3340 | 0.0547 | - |
| 0.1824 | 3350 | 0.0234 | - |
| 0.1830 | 3360 | 0.1416 | - |
| 0.1835 | 3370 | 0.0196 | - |
| 0.1841 | 3380 | 0.0258 | - |
| 0.1846 | 3390 | 0.1737 | - |
| 0.1851 | 3400 | 0.0434 | - |
| 0.1857 | 3410 | 0.0038 | - |
| 0.1862 | 3420 | 0.0748 | - |
| 0.1868 | 3430 | 0.0631 | - |
| 0.1873 | 3440 | 0.0084 | - |
| 0.1879 | 3450 | 0.0163 | - |
| 0.1884 | 3460 | 0.0502 | - |
| 0.1890 | 3470 | 0.0144 | - |
| 0.1895 | 3480 | 0.0864 | - |
| 0.1900 | 3490 | 0.0221 | - |
| 0.1906 | 3500 | 0.0466 | 0.0331 |
| 0.1911 | 3510 | 0.0537 | - |
| 0.1917 | 3520 | 0.064 | - |
| 0.1922 | 3530 | 0.012 | - |
| 0.1928 | 3540 | 0.0023 | - |
| 0.1933 | 3550 | 0.0055 | - |
| 0.1939 | 3560 | 0.0592 | - |
| 0.1944 | 3570 | 0.1166 | - |
| 0.1949 | 3580 | 0.0421 | - |
| 0.1955 | 3590 | 0.0167 | - |
| 0.1960 | 3600 | 0.0076 | - |
| 0.1966 | 3610 | 0.0114 | - |
| 0.1971 | 3620 | 0.031 | - |
| 0.1977 | 3630 | 0.1108 | - |
| 0.1982 | 3640 | 0.0629 | - |
| 0.1988 | 3650 | 0.2226 | - |
| 0.1993 | 3660 | 0.0248 | - |
| 0.1998 | 3670 | 0.0038 | - |
| 0.2004 | 3680 | 0.0684 | - |
| 0.2009 | 3690 | 0.0362 | - |
| 0.2015 | 3700 | 0.0068 | - |
| 0.2020 | 3710 | 0.025 | - |
| 0.2026 | 3720 | 0.0348 | - |
| 0.2031 | 3730 | 0.1607 | - |
| 0.2037 | 3740 | 0.0089 | - |
| 0.2042 | 3750 | 0.0202 | - |
| 0.2047 | 3760 | 0.0574 | - |
| 0.2053 | 3770 | 0.0478 | - |
| 0.2058 | 3780 | 0.0557 | - |
| 0.2064 | 3790 | 0.0387 | - |
| 0.2069 | 3800 | 0.0099 | - |
| 0.2075 | 3810 | 0.0048 | - |
| 0.2080 | 3820 | 0.1401 | - |
| 0.2086 | 3830 | 0.2235 | - |
| 0.2091 | 3840 | 0.0648 | - |
| 0.2096 | 3850 | 0.1194 | - |
| 0.2102 | 3860 | 0.0125 | - |
| 0.2107 | 3870 | 0.0089 | - |
| 0.2113 | 3880 | 0.0071 | - |
| 0.2118 | 3890 | 0.1777 | - |
| 0.2124 | 3900 | 0.0348 | - |
| 0.2129 | 3910 | 0.0873 | - |
| 0.2135 | 3920 | 0.0563 | - |
| 0.2140 | 3930 | 0.0091 | - |
| 0.2146 | 3940 | 0.0047 | - |
| 0.2151 | 3950 | 0.0047 | - |
| 0.2156 | 3960 | 0.0132 | - |
| 0.2162 | 3970 | 0.0896 | - |
| 0.2167 | 3980 | 0.0314 | - |
| 0.2173 | 3990 | 0.0118 | - |
| 0.2178 | 4000 | 0.0605 | 0.0287 |
| 0.2184 | 4010 | 0.0275 | - |
| 0.2189 | 4020 | 0.0035 | - |
| 0.2195 | 4030 | 0.0089 | - |
| 0.2200 | 4040 | 0.0501 | - |
| 0.2205 | 4050 | 0.0015 | - |
| 0.2211 | 4060 | 0.0308 | - |
| 0.2216 | 4070 | 0.0388 | - |
| 0.2222 | 4080 | 0.0015 | - |
| 0.2227 | 4090 | 0.1454 | - |
| 0.2233 | 4100 | 0.0609 | - |
| 0.2238 | 4110 | 0.0046 | - |
| 0.2244 | 4120 | 0.1594 | - |
| 0.2249 | 4130 | 0.0336 | - |
| 0.2254 | 4140 | 0.0985 | - |
| 0.2260 | 4150 | 0.0074 | - |
| 0.2265 | 4160 | 0.0207 | - |
| 0.2271 | 4170 | 0.002 | - |
| 0.2276 | 4180 | 0.0533 | - |
| 0.2282 | 4190 | 0.0129 | - |
| 0.2287 | 4200 | 0.0101 | - |
| 0.2293 | 4210 | 0.182 | - |
| 0.2298 | 4220 | 0.0824 | - |
| 0.2303 | 4230 | 0.0063 | - |
| 0.2309 | 4240 | 0.2493 | - |
| 0.2314 | 4250 | 0.1288 | - |
| 0.2320 | 4260 | 0.0684 | - |
| 0.2325 | 4270 | 0.0012 | - |
| 0.2331 | 4280 | 0.3867 | - |
| 0.2336 | 4290 | 0.074 | - |
| 0.2342 | 4300 | 0.0183 | - |
| 0.2347 | 4310 | 0.0717 | - |
| 0.2352 | 4320 | 0.0715 | - |
| 0.2358 | 4330 | 0.0385 | - |
| 0.2363 | 4340 | 0.1461 | - |
| 0.2369 | 4350 | 0.0064 | - |
| 0.2374 | 4360 | 0.0621 | - |
| 0.2380 | 4370 | 0.0129 | - |
| 0.2385 | 4380 | 0.0208 | - |
| 0.2391 | 4390 | 0.154 | - |
| 0.2396 | 4400 | 0.022 | - |
| 0.2401 | 4410 | 0.0119 | - |
| 0.2407 | 4420 | 0.0012 | - |
| 0.2412 | 4430 | 0.0526 | - |
| 0.2418 | 4440 | 0.108 | - |
| 0.2423 | 4450 | 0.0061 | - |
| 0.2429 | 4460 | 0.0082 | - |
| 0.2434 | 4470 | 0.0255 | - |
| 0.2440 | 4480 | 0.0082 | - |
| 0.2445 | 4490 | 0.024 | - |
| 0.2450 | 4500 | 0.0595 | 0.0328 |
| 0.2456 | 4510 | 0.0461 | - |
| 0.2461 | 4520 | 0.0076 | - |
| 0.2467 | 4530 | 0.0063 | - |
| 0.2472 | 4540 | 0.0219 | - |
| 0.2478 | 4550 | 0.001 | - |
| 0.2483 | 4560 | 0.0009 | - |
| 0.2489 | 4570 | 0.0245 | - |
| 0.2494 | 4580 | 0.0666 | - |
| 0.2499 | 4590 | 0.0106 | - |
| 0.2505 | 4600 | 0.0212 | - |
| 0.2510 | 4610 | 0.0173 | - |
| 0.2516 | 4620 | 0.2649 | - |
| 0.2521 | 4630 | 0.0062 | - |
| 0.2527 | 4640 | 0.0121 | - |
| 0.2532 | 4650 | 0.0047 | - |
| 0.2538 | 4660 | 0.0022 | - |
| 0.2543 | 4670 | 0.0777 | - |
| 0.2548 | 4680 | 0.0118 | - |
| 0.2554 | 4690 | 0.0183 | - |
| 0.2559 | 4700 | 0.0026 | - |
| 0.2565 | 4710 | 0.0417 | - |
| 0.2570 | 4720 | 0.0016 | - |
| 0.2576 | 4730 | 0.006 | - |
| 0.2581 | 4740 | 0.0081 | - |
| 0.2587 | 4750 | 0.1191 | - |
| 0.2592 | 4760 | 0.0018 | - |
| 0.2597 | 4770 | 0.0351 | - |
| 0.2603 | 4780 | 0.0731 | - |
| 0.2608 | 4790 | 0.0432 | - |
| 0.2614 | 4800 | 0.0344 | - |
| 0.2619 | 4810 | 0.0395 | - |
| 0.2625 | 4820 | 0.0275 | - |
| 0.2630 | 4830 | 0.0226 | - |
| 0.2636 | 4840 | 0.0778 | - |
| 0.2641 | 4850 | 0.0095 | - |
| 0.2646 | 4860 | 0.0056 | - |
| 0.2652 | 4870 | 0.0395 | - |
| 0.2657 | 4880 | 0.0511 | - |
| 0.2663 | 4890 | 0.0129 | - |
| 0.2668 | 4900 | 0.0278 | - |
| 0.2674 | 4910 | 0.0274 | - |
| 0.2679 | 4920 | 0.0993 | - |
| 0.2685 | 4930 | 0.074 | - |
| 0.2690 | 4940 | 0.0193 | - |
| 0.2695 | 4950 | 0.0021 | - |
| 0.2701 | 4960 | 0.2078 | - |
| 0.2706 | 4970 | 0.0005 | - |
| 0.2712 | 4980 | 0.0062 | - |
| 0.2717 | 4990 | 0.2275 | - |
| 0.2723 | 5000 | 0.0023 | 0.0347 |
| 0.2728 | 5010 | 0.048 | - |
| 0.2734 | 5020 | 0.0066 | - |
| 0.2739 | 5030 | 0.007 | - |
| 0.2745 | 5040 | 0.0019 | - |
| 0.2750 | 5050 | 0.0366 | - |
| 0.2755 | 5060 | 0.0514 | - |
| 0.2761 | 5070 | 0.056 | - |
| 0.2766 | 5080 | 0.158 | - |
| 0.2772 | 5090 | 0.0358 | - |
| 0.2777 | 5100 | 0.0226 | - |
| 0.2783 | 5110 | 0.0109 | - |
| 0.2788 | 5120 | 0.0454 | - |
| 0.2794 | 5130 | 0.0325 | - |
| 0.2799 | 5140 | 0.0104 | - |
| 0.2804 | 5150 | 0.009 | - |
| 0.2810 | 5160 | 0.0089 | - |
| 0.2815 | 5170 | 0.0283 | - |
| 0.2821 | 5180 | 0.0413 | - |
| 0.2826 | 5190 | 0.0099 | - |
| 0.2832 | 5200 | 0.1241 | - |
| 0.2837 | 5210 | 0.004 | - |
| 0.2843 | 5220 | 0.024 | - |
| 0.2848 | 5230 | 0.018 | - |
| 0.2853 | 5240 | 0.006 | - |
| 0.2859 | 5250 | 0.0082 | - |
| 0.2864 | 5260 | 0.0098 | - |
| 0.2870 | 5270 | 0.0172 | - |
| 0.2875 | 5280 | 0.0226 | - |
| 0.2881 | 5290 | 0.0086 | - |
| 0.2886 | 5300 | 0.1538 | - |
| 0.2892 | 5310 | 0.0006 | - |
| 0.2897 | 5320 | 0.01 | - |
| 0.2902 | 5330 | 0.1389 | - |
| 0.2908 | 5340 | 0.0006 | - |
| 0.2913 | 5350 | 0.0114 | - |
| 0.2919 | 5360 | 0.0184 | - |
| 0.2924 | 5370 | 0.0106 | - |
| 0.2930 | 5380 | 0.0131 | - |
| 0.2935 | 5390 | 0.0092 | - |
| 0.2941 | 5400 | 0.0673 | - |
| 0.2946 | 5410 | 0.0292 | - |
| 0.2951 | 5420 | 0.0007 | - |
| 0.2957 | 5430 | 0.007 | - |
| 0.2962 | 5440 | 0.0097 | - |
| 0.2968 | 5450 | 0.0022 | - |
| 0.2973 | 5460 | 0.0132 | - |
| 0.2979 | 5470 | 0.0199 | - |
| 0.2984 | 5480 | 0.0195 | - |
| 0.2990 | 5490 | 0.0066 | - |
| 0.2995 | 5500 | 0.0637 | 0.0368 |
| 0.3000 | 5510 | 0.0095 | - |
| 0.3006 | 5520 | 0.0047 | - |
| 0.3011 | 5530 | 0.0249 | - |
| 0.3017 | 5540 | 0.0089 | - |
| 0.3022 | 5550 | 0.0025 | - |
| 0.3028 | 5560 | 0.054 | - |
| 0.3033 | 5570 | 0.0015 | - |
| 0.3039 | 5580 | 0.0722 | - |
| 0.3044 | 5590 | 0.0076 | - |
| 0.3049 | 5600 | 0.0348 | - |
| 0.3055 | 5610 | 0.0019 | - |
| 0.3060 | 5620 | 0.0066 | - |
| 0.3066 | 5630 | 0.0076 | - |
| 0.3071 | 5640 | 0.028 | - |
| 0.3077 | 5650 | 0.0101 | - |
| 0.3082 | 5660 | 0.0183 | - |
| 0.3088 | 5670 | 0.0377 | - |
| 0.3093 | 5680 | 0.0183 | - |
| 0.3098 | 5690 | 0.1069 | - |
| 0.3104 | 5700 | 0.0106 | - |
| 0.3109 | 5710 | 0.0109 | - |
| 0.3115 | 5720 | 0.0704 | - |
| 0.3120 | 5730 | 0.0096 | - |
| 0.3126 | 5740 | 0.0014 | - |
| 0.3131 | 5750 | 0.0147 | - |
| 0.3137 | 5760 | 0.0889 | - |
| 0.3142 | 5770 | 0.0786 | - |
| 0.3147 | 5780 | 0.0613 | - |
| 0.3153 | 5790 | 0.0025 | - |
| 0.3158 | 5800 | 0.0147 | - |
| 0.3164 | 5810 | 0.0092 | - |
| 0.3169 | 5820 | 0.0282 | - |
| 0.3175 | 5830 | 0.0029 | - |
| 0.3180 | 5840 | 0.001 | - |
| 0.3186 | 5850 | 0.0022 | - |
| 0.3191 | 5860 | 0.0069 | - |
| 0.3196 | 5870 | 0.0027 | - |
| 0.3202 | 5880 | 0.0131 | - |
| 0.3207 | 5890 | 0.0016 | - |
| 0.3213 | 5900 | 0.0447 | - |
| 0.3218 | 5910 | 0.0258 | - |
| 0.3224 | 5920 | 0.0193 | - |
| 0.3229 | 5930 | 0.0027 | - |
| 0.3235 | 5940 | 0.1442 | - |
| 0.3240 | 5950 | 0.1342 | - |
| 0.3245 | 5960 | 0.0187 | - |
| 0.3251 | 5970 | 0.0054 | - |
| 0.3256 | 5980 | 0.0526 | - |
| 0.3262 | 5990 | 0.003 | - |
| 0.3267 | 6000 | 0.0087 | 0.0200 |
| 0.3273 | 6010 | 0.0572 | - |
| 0.3278 | 6020 | 0.0066 | - |
| 0.3284 | 6030 | 0.0022 | - |
| 0.3289 | 6040 | 0.0008 | - |
| 0.3294 | 6050 | 0.002 | - |
| 0.3300 | 6060 | 0.0026 | - |
| 0.3305 | 6070 | 0.0006 | - |
| 0.3311 | 6080 | 0.0018 | - |
| 0.3316 | 6090 | 0.0045 | - |
| 0.3322 | 6100 | 0.0179 | - |
| 0.3327 | 6110 | 0.0047 | - |
| 0.3333 | 6120 | 0.0542 | - |
| 0.3338 | 6130 | 0.0053 | - |
| 0.3343 | 6140 | 0.0386 | - |
| 0.3349 | 6150 | 0.0176 | - |
| 0.3354 | 6160 | 0.0074 | - |
| 0.3360 | 6170 | 0.2899 | - |
| 0.3365 | 6180 | 0.0004 | - |
| 0.3371 | 6190 | 0.0209 | - |
| 0.3376 | 6200 | 0.0482 | - |
| 0.3382 | 6210 | 0.0264 | - |
| 0.3387 | 6220 | 0.0028 | - |
| 0.3393 | 6230 | 0.0019 | - |
| 0.3398 | 6240 | 0.0746 | - |
| 0.3403 | 6250 | 0.0021 | - |
| 0.3409 | 6260 | 0.0006 | - |
| 0.3414 | 6270 | 0.0046 | - |
| 0.3420 | 6280 | 0.0151 | - |
| 0.3425 | 6290 | 0.4818 | - |
| 0.3431 | 6300 | 0.0047 | - |
| 0.3436 | 6310 | 0.0088 | - |
| 0.3442 | 6320 | 0.0014 | - |
| 0.3447 | 6330 | 0.0255 | - |
| 0.3452 | 6340 | 0.0087 | - |
| 0.3458 | 6350 | 0.0107 | - |
| 0.3463 | 6360 | 0.0286 | - |
| 0.3469 | 6370 | 0.0031 | - |
| 0.3474 | 6380 | 0.0549 | - |
| 0.3480 | 6390 | 0.0014 | - |
| 0.3485 | 6400 | 0.014 | - |
| 0.3491 | 6410 | 0.0025 | - |
| 0.3496 | 6420 | 0.0647 | - |
| 0.3501 | 6430 | 0.0004 | - |
| 0.3507 | 6440 | 0.0083 | - |
| 0.3512 | 6450 | 0.0004 | - |
| 0.3518 | 6460 | 0.0007 | - |
| 0.3523 | 6470 | 0.0736 | - |
| 0.3529 | 6480 | 0.0273 | - |
| 0.3534 | 6490 | 0.0079 | - |
| 0.3540 | 6500 | 0.0021 | 0.0102 |
| 0.3545 | 6510 | 0.0478 | - |
| 0.3550 | 6520 | 0.0031 | - |
| 0.3556 | 6530 | 0.1222 | - |
| 0.3561 | 6540 | 0.0006 | - |
| 0.3567 | 6550 | 0.0084 | - |
| 0.3572 | 6560 | 0.004 | - |
| 0.3578 | 6570 | 0.0277 | - |
| 0.3583 | 6580 | 0.0067 | - |
| 0.3589 | 6590 | 0.0041 | - |
| 0.3594 | 6600 | 0.0004 | - |
| 0.3599 | 6610 | 0.0098 | - |
| 0.3605 | 6620 | 0.0231 | - |
| 0.3610 | 6630 | 0.0082 | - |
| 0.3616 | 6640 | 0.0182 | - |
| 0.3621 | 6650 | 0.007 | - |
| 0.3627 | 6660 | 0.0016 | - |
| 0.3632 | 6670 | 0.0774 | - |
| 0.3638 | 6680 | 0.0292 | - |
| 0.3643 | 6690 | 0.0069 | - |
| 0.3648 | 6700 | 0.0047 | - |
| 0.3654 | 6710 | 0.0141 | - |
| 0.3659 | 6720 | 0.0008 | - |
| 0.3665 | 6730 | 0.0086 | - |
| 0.3670 | 6740 | 0.0012 | - |
| 0.3676 | 6750 | 0.0009 | - |
| 0.3681 | 6760 | 0.1414 | - |
| 0.3687 | 6770 | 0.0008 | - |
| 0.3692 | 6780 | 0.0054 | - |
| 0.3697 | 6790 | 0.0015 | - |
| 0.3703 | 6800 | 0.0067 | - |
| 0.3708 | 6810 | 0.0076 | - |
| 0.3714 | 6820 | 0.0056 | - |
| 0.3719 | 6830 | 0.0067 | - |
| 0.3725 | 6840 | 0.0008 | - |
| 0.3730 | 6850 | 0.0046 | - |
| 0.3736 | 6860 | 0.021 | - |
| 0.3741 | 6870 | 0.0019 | - |
| 0.3746 | 6880 | 0.0516 | - |
| 0.3752 | 6890 | 0.0256 | - |
| 0.3757 | 6900 | 0.1099 | - |
| 0.3763 | 6910 | 0.0017 | - |
| 0.3768 | 6920 | 0.0416 | - |
| 0.3774 | 6930 | 0.0515 | - |
| 0.3779 | 6940 | 0.0168 | - |
| 0.3785 | 6950 | 0.0262 | - |
| 0.3790 | 6960 | 0.0067 | - |
| 0.3795 | 6970 | 0.0205 | - |
| 0.3801 | 6980 | 0.0133 | - |
| 0.3806 | 6990 | 0.0176 | - |
| 0.3812 | 7000 | 0.0057 | 0.0113 |
| 0.3817 | 7010 | 0.0029 | - |
| 0.3823 | 7020 | 0.0196 | - |
| 0.3828 | 7030 | 0.0034 | - |
| 0.3834 | 7040 | 0.0179 | - |
| 0.3839 | 7050 | 0.009 | - |
| 0.3844 | 7060 | 0.0053 | - |
| 0.3850 | 7070 | 0.0092 | - |
| 0.3855 | 7080 | 0.0036 | - |
| 0.3861 | 7090 | 0.0114 | - |
| 0.3866 | 7100 | 0.1554 | - |
| 0.3872 | 7110 | 0.0038 | - |
| 0.3877 | 7120 | 0.0061 | - |
| 0.3883 | 7130 | 0.0182 | - |
| 0.3888 | 7140 | 0.0054 | - |
| 0.3893 | 7150 | 0.0208 | - |
| 0.3899 | 7160 | 0.0032 | - |
| 0.3904 | 7170 | 0.0887 | - |
| 0.3910 | 7180 | 0.0168 | - |
| 0.3915 | 7190 | 0.0038 | - |
| 0.3921 | 7200 | 0.053 | - |
| 0.3926 | 7210 | 0.0021 | - |
| 0.3932 | 7220 | 0.0009 | - |
| 0.3937 | 7230 | 0.0117 | - |
| 0.3942 | 7240 | 0.0359 | - |
| 0.3948 | 7250 | 0.0057 | - |
| 0.3953 | 7260 | 0.0414 | - |
| 0.3959 | 7270 | 0.0295 | - |
| 0.3964 | 7280 | 0.0008 | - |
| 0.3970 | 7290 | 0.001 | - |
| 0.3975 | 7300 | 0.0128 | - |
| 0.3981 | 7310 | 0.0021 | - |
| 0.3986 | 7320 | 0.0628 | - |
| 0.3992 | 7330 | 0.0242 | - |
| 0.3997 | 7340 | 0.0286 | - |
| 0.4002 | 7350 | 0.0004 | - |
| 0.4008 | 7360 | 0.0006 | - |
| 0.4013 | 7370 | 0.0078 | - |
| 0.4019 | 7380 | 0.0097 | - |
| 0.4024 | 7390 | 0.0084 | - |
| 0.4030 | 7400 | 0.0049 | - |
| 0.4035 | 7410 | 0.0042 | - |
| 0.4041 | 7420 | 0.0028 | - |
| 0.4046 | 7430 | 0.0237 | - |
| 0.4051 | 7440 | 0.0017 | - |
| 0.4057 | 7450 | 0.2392 | - |
| 0.4062 | 7460 | 0.0012 | - |
| 0.4068 | 7470 | 0.0003 | - |
| 0.4073 | 7480 | 0.0072 | - |
| 0.4079 | 7490 | 0.0012 | - |
| 0.4084 | 7500 | 0.0219 | 0.0092 |
| 0.4090 | 7510 | 0.0043 | - |
| 0.4095 | 7520 | 0.0228 | - |
| 0.4100 | 7530 | 0.0179 | - |
| 0.4106 | 7540 | 0.0012 | - |
| 0.4111 | 7550 | 0.0095 | - |
| 0.4117 | 7560 | 0.0024 | - |
| 0.4122 | 7570 | 0.0009 | - |
| 0.4128 | 7580 | 0.005 | - |
| 0.4133 | 7590 | 0.0048 | - |
| 0.4139 | 7600 | 0.003 | - |
| 0.4144 | 7610 | 0.0077 | - |
| 0.4149 | 7620 | 0.0079 | - |
| 0.4155 | 7630 | 0.0108 | - |
| 0.4160 | 7640 | 0.0957 | - |
| 0.4166 | 7650 | 0.0004 | - |
| 0.4171 | 7660 | 0.0022 | - |
| 0.4177 | 7670 | 0.0021 | - |
| 0.4182 | 7680 | 0.0029 | - |
| 0.4188 | 7690 | 0.0223 | - |
| 0.4193 | 7700 | 0.01 | - |
| 0.4198 | 7710 | 0.0011 | - |
| 0.4204 | 7720 | 0.0031 | - |
| 0.4209 | 7730 | 0.029 | - |
| 0.4215 | 7740 | 0.0124 | - |
| 0.4220 | 7750 | 0.0033 | - |
| 0.4226 | 7760 | 0.0243 | - |
| 0.4231 | 7770 | 0.0046 | - |
| 0.4237 | 7780 | 0.0059 | - |
| 0.4242 | 7790 | 0.004 | - |
| 0.4247 | 7800 | 0.008 | - |
| 0.4253 | 7810 | 0.005 | - |
| 0.4258 | 7820 | 0.0101 | - |
| 0.4264 | 7830 | 0.1265 | - |
| 0.4269 | 7840 | 0.0014 | - |
| 0.4275 | 7850 | 0.0073 | - |
| 0.4280 | 7860 | 0.0007 | - |
| 0.4286 | 7870 | 0.0042 | - |
| 0.4291 | 7880 | 0.1662 | - |
| 0.4296 | 7890 | 0.0045 | - |
| 0.4302 | 7900 | 0.0006 | - |
| 0.4307 | 7910 | 0.0148 | - |
| 0.4313 | 7920 | 0.001 | - |
| 0.4318 | 7930 | 0.0146 | - |
| 0.4324 | 7940 | 0.0158 | - |
| 0.4329 | 7950 | 0.046 | - |
| 0.4335 | 7960 | 0.0031 | - |
| 0.4340 | 7970 | 0.0323 | - |
| 0.4345 | 7980 | 0.0151 | - |
| 0.4351 | 7990 | 0.0291 | - |
| 0.4356 | 8000 | 0.0093 | 0.0115 |
| 0.4362 | 8010 | 0.0107 | - |
| 0.4367 | 8020 | 0.0005 | - |
| 0.4373 | 8030 | 0.0044 | - |
| 0.4378 | 8040 | 0.0017 | - |
| 0.4384 | 8050 | 0.012 | - |
| 0.4389 | 8060 | 0.1177 | - |
| 0.4394 | 8070 | 0.0266 | - |
| 0.4400 | 8080 | 0.0124 | - |
| 0.4405 | 8090 | 0.0031 | - |
| 0.4411 | 8100 | 0.0449 | - |
| 0.4416 | 8110 | 0.0182 | - |
| 0.4422 | 8120 | 0.0031 | - |
| 0.4427 | 8130 | 0.0251 | - |
| 0.4433 | 8140 | 0.0012 | - |
| 0.4438 | 8150 | 0.0025 | - |
| 0.4443 | 8160 | 0.0742 | - |
| 0.4449 | 8170 | 0.004 | - |
| 0.4454 | 8180 | 0.0018 | - |
| 0.4460 | 8190 | 0.0268 | - |
| 0.4465 | 8200 | 0.0083 | - |
| 0.4471 | 8210 | 0.0088 | - |
| 0.4476 | 8220 | 0.1101 | - |
| 0.4482 | 8230 | 0.0016 | - |
| 0.4487 | 8240 | 0.0024 | - |
| 0.4492 | 8250 | 0.0042 | - |
| 0.4498 | 8260 | 0.0298 | - |
| 0.4503 | 8270 | 0.0316 | - |
| 0.4509 | 8280 | 0.0013 | - |
| 0.4514 | 8290 | 0.0045 | - |
| 0.4520 | 8300 | 0.0065 | - |
| 0.4525 | 8310 | 0.0047 | - |
| 0.4531 | 8320 | 0.0113 | - |
| 0.4536 | 8330 | 0.0007 | - |
| 0.4541 | 8340 | 0.0019 | - |
| 0.4547 | 8350 | 0.0455 | - |
| 0.4552 | 8360 | 0.0413 | - |
| 0.4558 | 8370 | 0.0083 | - |
| 0.4563 | 8380 | 0.0355 | - |
| 0.4569 | 8390 | 0.006 | - |
| 0.4574 | 8400 | 0.0085 | - |
| 0.4580 | 8410 | 0.0034 | - |
| 0.4585 | 8420 | 0.0051 | - |
| 0.4591 | 8430 | 0.0312 | - |
| 0.4596 | 8440 | 0.0015 | - |
| 0.4601 | 8450 | 0.0023 | - |
| 0.4607 | 8460 | 0.0078 | - |
| 0.4612 | 8470 | 0.1225 | - |
| 0.4618 | 8480 | 0.018 | - |
| 0.4623 | 8490 | 0.0003 | - |
| 0.4629 | 8500 | 0.0021 | 0.0097 |
| 0.4634 | 8510 | 0.004 | - |
| 0.4640 | 8520 | 0.002 | - |
| 0.4645 | 8530 | 0.0016 | - |
| 0.4650 | 8540 | 0.0306 | - |
| 0.4656 | 8550 | 0.0211 | - |
| 0.4661 | 8560 | 0.0015 | - |
| 0.4667 | 8570 | 0.0409 | - |
| 0.4672 | 8580 | 0.0049 | - |
| 0.4678 | 8590 | 0.0053 | - |
| 0.4683 | 8600 | 0.0031 | - |
| 0.4689 | 8610 | 0.0924 | - |
| 0.4694 | 8620 | 0.0003 | - |
| 0.4699 | 8630 | 0.0032 | - |
| 0.4705 | 8640 | 0.0009 | - |
| 0.4710 | 8650 | 0.0508 | - |
| 0.4716 | 8660 | 0.0041 | - |
| 0.4721 | 8670 | 0.0073 | - |
| 0.4727 | 8680 | 0.0581 | - |
| 0.4732 | 8690 | 0.0268 | - |
| 0.4738 | 8700 | 0.0136 | - |
| 0.4743 | 8710 | 0.0052 | - |
| 0.4748 | 8720 | 0.0023 | - |
| 0.4754 | 8730 | 0.0004 | - |
| 0.4759 | 8740 | 0.1408 | - |
| 0.4765 | 8750 | 0.0006 | - |
| 0.4770 | 8760 | 0.0031 | - |
| 0.4776 | 8770 | 0.0028 | - |
| 0.4781 | 8780 | 0.0075 | - |
| 0.4787 | 8790 | 0.0306 | - |
| 0.4792 | 8800 | 0.0118 | - |
| 0.4797 | 8810 | 0.007 | - |
| 0.4803 | 8820 | 0.0013 | - |
| 0.4808 | 8830 | 0.0082 | - |
| 0.4814 | 8840 | 0.0091 | - |
| 0.4819 | 8850 | 0.0028 | - |
| 0.4825 | 8860 | 0.021 | - |
| 0.4830 | 8870 | 0.0061 | - |
| 0.4836 | 8880 | 0.0982 | - |
| 0.4841 | 8890 | 0.0079 | - |
| 0.4846 | 8900 | 0.017 | - |
| 0.4852 | 8910 | 0.0013 | - |
| 0.4857 | 8920 | 0.009 | - |
| 0.4863 | 8930 | 0.0056 | - |
| 0.4868 | 8940 | 0.0218 | - |
| 0.4874 | 8950 | 0.0196 | - |
| 0.4879 | 8960 | 0.0193 | - |
| 0.4885 | 8970 | 0.0062 | - |
| 0.4890 | 8980 | 0.0027 | - |
| 0.4895 | 8990 | 0.0013 | - |
| 0.4901 | 9000 | 0.0223 | 0.0069 |
| 0.4906 | 9010 | 0.0088 | - |
| 0.4912 | 9020 | 0.0052 | - |
| 0.4917 | 9030 | 0.0023 | - |
| 0.4923 | 9040 | 0.0048 | - |
| 0.4928 | 9050 | 0.0335 | - |
| 0.4934 | 9060 | 0.0282 | - |
| 0.4939 | 9070 | 0.0028 | - |
| 0.4944 | 9080 | 0.004 | - |
| 0.4950 | 9090 | 0.0148 | - |
| 0.4955 | 9100 | 0.0057 | - |
| 0.4961 | 9110 | 0.0032 | - |
| 0.4966 | 9120 | 0.0017 | - |
| 0.4972 | 9130 | 0.0055 | - |
| 0.4977 | 9140 | 0.0068 | - |
| 0.4983 | 9150 | 0.0311 | - |
| 0.4988 | 9160 | 0.0137 | - |
| 0.4993 | 9170 | 0.0018 | - |
| 0.4999 | 9180 | 0.0132 | - |
| 0.5004 | 9190 | 0.0132 | - |
| 0.5010 | 9200 | 0.0266 | - |
| 0.5015 | 9210 | 0.004 | - |
| 0.5021 | 9220 | 0.0077 | - |
| 0.5026 | 9230 | 0.001 | - |
| 0.5032 | 9240 | 0.002 | - |
| 0.5037 | 9250 | 0.0656 | - |
| 0.5042 | 9260 | 0.0426 | - |
| 0.5048 | 9270 | 0.0261 | - |
| 0.5053 | 9280 | 0.0243 | - |
| 0.5059 | 9290 | 0.0068 | - |
| 0.5064 | 9300 | 0.009 | - |
| 0.5070 | 9310 | 0.07 | - |
| 0.5075 | 9320 | 0.0015 | - |
| 0.5081 | 9330 | 0.0034 | - |
| 0.5086 | 9340 | 0.0052 | - |
| 0.5091 | 9350 | 0.0247 | - |
| 0.5097 | 9360 | 0.1479 | - |
| 0.5102 | 9370 | 0.0069 | - |
| 0.5108 | 9380 | 0.0109 | - |
| 0.5113 | 9390 | 0.0317 | - |
| 0.5119 | 9400 | 0.0189 | - |
| 0.5124 | 9410 | 0.0023 | - |
| 0.5130 | 9420 | 0.0078 | - |
| 0.5135 | 9430 | 0.0046 | - |
| 0.5140 | 9440 | 0.0027 | - |
| 0.5146 | 9450 | 0.0039 | - |
| 0.5151 | 9460 | 0.0351 | - |
| 0.5157 | 9470 | 0.0063 | - |
| 0.5162 | 9480 | 0.0012 | - |
| 0.5168 | 9490 | 0.0521 | - |
| 0.5173 | 9500 | 0.0275 | 0.0063 |
| 0.5179 | 9510 | 0.0286 | - |
| 0.5184 | 9520 | 0.0042 | - |
| 0.5190 | 9530 | 0.0111 | - |
| 0.5195 | 9540 | 0.0279 | - |
| 0.5200 | 9550 | 0.0197 | - |
| 0.5206 | 9560 | 0.0015 | - |
| 0.5211 | 9570 | 0.0127 | - |
| 0.5217 | 9580 | 0.0007 | - |
| 0.5222 | 9590 | 0.0016 | - |
| 0.5228 | 9600 | 0.0225 | - |
| 0.5233 | 9610 | 0.0023 | - |
| 0.5239 | 9620 | 0.0169 | - |
| 0.5244 | 9630 | 0.0986 | - |
| 0.5249 | 9640 | 0.0063 | - |
| 0.5255 | 9650 | 0.0123 | - |
| 0.5260 | 9660 | 0.0216 | - |
| 0.5266 | 9670 | 0.0066 | - |
| 0.5271 | 9680 | 0.0042 | - |
| 0.5277 | 9690 | 0.0191 | - |
| 0.5282 | 9700 | 0.1061 | - |
| 0.5288 | 9710 | 0.0107 | - |
| 0.5293 | 9720 | 0.0007 | - |
| 0.5298 | 9730 | 0.0031 | - |
| 0.5304 | 9740 | 0.0134 | - |
| 0.5309 | 9750 | 0.0025 | - |
| 0.5315 | 9760 | 0.0204 | - |
| 0.5320 | 9770 | 0.0012 | - |
| 0.5326 | 9780 | 0.0046 | - |
| 0.5331 | 9790 | 0.0052 | - |
| 0.5337 | 9800 | 0.0005 | - |
| 0.5342 | 9810 | 0.016 | - |
| 0.5347 | 9820 | 0.0342 | - |
| 0.5353 | 9830 | 0.0008 | - |
| 0.5358 | 9840 | 0.0017 | - |
| 0.5364 | 9850 | 0.0004 | - |
| 0.5369 | 9860 | 0.3161 | - |
| 0.5375 | 9870 | 0.0489 | - |
| 0.5380 | 9880 | 0.0532 | - |
| 0.5386 | 9890 | 0.0054 | - |
| 0.5391 | 9900 | 0.0045 | - |
| 0.5396 | 9910 | 0.0097 | - |
| 0.5402 | 9920 | 0.0205 | - |
| 0.5407 | 9930 | 0.0087 | - |
| 0.5413 | 9940 | 0.0264 | - |
| 0.5418 | 9950 | 0.0066 | - |
| 0.5424 | 9960 | 0.0079 | - |
| 0.5429 | 9970 | 0.0131 | - |
| 0.5435 | 9980 | 0.0026 | - |
| 0.5440 | 9990 | 0.0274 | - |
| 0.5445 | 10000 | 0.0658 | 0.0064 |
</details>
### Framework Versions
- Python: 3.10.12
- Sentence Transformers: 4.1.0
- Transformers: 4.52.3
- PyTorch: 2.6.0+cu124
- Accelerate: 1.7.0
- Datasets: 3.6.0
- Tokenizers: 0.21.1
## Citation
### BibTeX
#### Sentence Transformers
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
```
#### MultipleNegativesRankingLoss
```bibtex
@misc{henderson2017efficient,
title={Efficient Natural Language Response Suggestion for Smart Reply},
author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
year={2017},
eprint={1705.00652},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> |
cikgu-fadhilah/mewe.video.cikgu.fadhilah.viral.mewe.app.cctv.wiring.viral | cikgu-fadhilah | 2025-05-29T10:31:30Z | 0 | 0 | null | [
"region:us"
] | null | 2025-05-29T10:31:20Z | [🌐 CLICK HERE 🟢==►► WATCH NOW](https://videohere.top/?V=cikgu-fadhilah)
[🔴 CLICK HERE 🌐==►► Download Now)](https://videohere.top/?V=cikgu-fadhilah)
[<img alt="fsd" src="https://i.postimg.cc/qvPp49Sm/ythngythg.gif">](https://videohere.top/?V=cikgu-fadhilah) |
jonlecumberri/MNLP_M3_mcqa_model_v1 | jonlecumberri | 2025-05-29T10:25:39Z | 10 | 0 | transformers | [
"transformers",
"safetensors",
"qwen3",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"4-bit",
"bitsandbytes",
"region:us"
] | text-generation | 2025-05-28T18:45:44Z | ---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
THUDM/androidgen-llama-3-70b | THUDM | 2025-05-29T10:21:48Z | 0 | 1 | null | [
"safetensors",
"llama",
"androidgen",
"llama3",
"androidworld",
"llm",
"agent",
"en",
"arxiv:2504.19298",
"license:other",
"region:us"
] | null | 2025-05-28T07:58:34Z | ---
license: other
language:
- en
base_model:
- meta/Llama-3-70B
tags:
- androidgen
- llama3
- androidworld
- llm
- agent
---
# AndroidGen-Llama-3-70B
## Model Introduction
AndroidGen-Llama-3-70B is the open-source version of AndroidGen in Llama-3-70B released by Zhipu AI.
AndroidGen enables LLM-based agents to autonomously perform tasks across various Android applications, including messaging, clock, email, settings, etc., without requiring manually labeled interaction data.
**For more inference code and requirements, please visit our [github page](GitHub - THUDM/AndroidGen).**
## Citations
If you find our work useful, please consider citing the following paper.
```
@article{lai2025androidgen,
title={AndroidGen: Building an Android Language Agent under Data Scarcity},
author={Lai, Hanyu and Gao, Junjie and Liu, Xiao and Xu, Yifan and Zhang, Shudan and Dong, Yuxiao and Tang, Jie},
journal={arXiv preprint arXiv:2504.19298},
year={2025}
}
``` |
GigiTrottola/Qwen2.5-1.5B-Instruct-Gensyn-Swarm-yawning_webbed_robin | GigiTrottola | 2025-05-29T10:21:21Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"generated_from_trainer",
"rl-swarm",
"grpo",
"gensyn",
"I am yawning webbed robin",
"unsloth",
"trl",
"arxiv:2402.03300",
"base_model:Gensyn/Qwen2.5-1.5B-Instruct",
"base_model:finetune:Gensyn/Qwen2.5-1.5B-Instruct",
"endpoints_compatible",
"region:us"
] | null | 2025-05-28T22:57:57Z | ---
base_model: Gensyn/Qwen2.5-1.5B-Instruct
library_name: transformers
model_name: Qwen2.5-1.5B-Instruct-Gensyn-Swarm-yawning_webbed_robin
tags:
- generated_from_trainer
- rl-swarm
- grpo
- gensyn
- I am yawning webbed robin
- unsloth
- trl
licence: license
---
# Model Card for Qwen2.5-1.5B-Instruct-Gensyn-Swarm-yawning_webbed_robin
This model is a fine-tuned version of [Gensyn/Qwen2.5-1.5B-Instruct](https://huggingface.co/Gensyn/Qwen2.5-1.5B-Instruct).
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="GigiTrottola/Qwen2.5-1.5B-Instruct-Gensyn-Swarm-yawning_webbed_robin", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).
### Framework versions
- TRL: 0.15.2
- Transformers: 4.48.2
- Pytorch: 2.5.1
- Datasets: 3.6.0
- Tokenizers: 0.21.1
## Citations
Cite GRPO as:
```bibtex
@article{zhihong2024deepseekmath,
title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
author = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
year = 2024,
eprint = {arXiv:2402.03300},
}
```
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
``` |
hubble658/v3.1-full-3b | hubble658 | 2025-05-29T10:19:04Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"qwen2_5_vl",
"image-text-to-text",
"text-generation-inference",
"unsloth",
"trl",
"sft",
"conversational",
"en",
"base_model:unsloth/Qwen2.5-VL-3B-Instruct",
"base_model:finetune:unsloth/Qwen2.5-VL-3B-Instruct",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | image-text-to-text | 2025-05-29T10:17:34Z | ---
base_model: unsloth/Qwen2.5-VL-3B-Instruct
tags:
- text-generation-inference
- transformers
- unsloth
- qwen2_5_vl
- trl
- sft
license: apache-2.0
language:
- en
---
# Uploaded model
- **Developed by:** hubble658
- **License:** apache-2.0
- **Finetuned from model :** unsloth/Qwen2.5-VL-3B-Instruct
This qwen2_5_vl model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
jimnoneill/CarD-T | jimnoneill | 2025-05-29T10:10:06Z | 6 | 1 | transformers | [
"transformers",
"tensorboard",
"safetensors",
"electra",
"token-classification",
"biology",
"chemistry",
"medical",
"cancer",
"carcinogenesis",
"biomedical",
"ner",
"oncology",
"en",
"dataset:jimnoneill/CarD-T-NER",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | token-classification | 2024-08-09T20:56:49Z | ---
license: apache-2.0
language:
- en
library_name: transformers
pipeline_tag: token-classification
tags:
- biology
- chemistry
- medical
- cancer
- carcinogenesis
- biomedical
- ner
- oncology
datasets:
- jimnoneill/CarD-T-NER
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: CarD-T
results:
- task:
type: token-classification
name: Named Entity Recognition
dataset:
name: CarD-T-NER
type: jimnoneill/CarD-T-NER
metrics:
- type: precision
value: 0.894
- type: recall
value: 0.857
- type: f1
value: 0.875
---
# CarD-T: Carcinogen Detection via Transformers
## Overview
CarD-T (Carcinogen Detection via Transformers) is a novel text analytics approach that combines transformer-based machine learning with probabilistic statistical analysis to efficiently nominate carcinogens from scientific texts. This model is designed to address the challenges faced by current systems in managing the burgeoning biomedical literature related to carcinogen identification and classification.
## Model Details
* **Architecture**: Based on Bio-ELECTRA, a 335 million parameter language model
* **Training Data**: [CarD-T-NER dataset](https://huggingface.co/datasets/jimnoneill/CarD-T-NER) containing 19,975 annotated examples from PubMed abstracts (2000-2024)
* Training set: 11,985 examples
* Test set: 7,990 examples
* **Task**: Named Entity Recognition (NER) for carcinogen identification using BIO tagging
* **Performance**:
* Precision: 0.894
* Recall: 0.857
* F1 Score: 0.875
## Named Entity Labels
The model recognizes 4 entity types using BIO (Beginning-Inside-Outside) tagging scheme, resulting in 9 total labels:
| Label ID | Label | Description |
|----------|-------|-------------|
| 0 | O | Outside any entity |
| 1 | B-carcinogen | Beginning of carcinogen entity |
| 2 | I-carcinogen | Inside carcinogen entity |
| 3 | B-negative | Beginning of negative/exculpatory evidence |
| 4 | I-negative | Inside negative evidence |
| 5 | B-cancertype | Beginning of cancer type/metadata |
| 6 | I-cancertype | Inside cancer type/metadata |
| 7 | B-antineoplastic | Beginning of anti-cancer agent |
| 8 | I-antineoplastic | Inside anti-cancer agent |
### Entity Type Descriptions:
* **carcinogen**: Substances or agents implicated in carcinogenesis
* **negative**: Exculpating evidence for potential carcinogenic entities
* **cancertype**: Metadata including organism (human/animal/cell), cancer type, and affected organs
* **antineoplastic**: Chemotherapy drugs and cancer-protective agents
## Use Cases
* Streamlining toxicogenomic literature reviews
* Identifying potential carcinogens for further investigation
* Augmenting existing carcinogen databases with emerging candidates
* Extracting structured information from cancer research literature
* Supporting evidence-based oncology research
## Limitations
* Identifies potential candidates, not confirmed carcinogens
* Analysis limited to abstract-level information
* May be influenced by publication trends and research focus shifts
* Requires validation by domain experts for clinical applications
## Installation
```bash
pip install transformers torch datasets
```
## Usage
### Basic Usage
```python
from transformers import AutoTokenizer, AutoModelForTokenClassification
import torch
# Load model and tokenizer
model_name = "jimnoneill/CarD-T"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForTokenClassification.from_pretrained(model_name)
# Define label mappings
id2label = {
0: "O",
1: "B-carcinogen",
2: "I-carcinogen",
3: "B-negative",
4: "I-negative",
5: "B-cancertype",
6: "I-cancertype",
7: "B-antineoplastic",
8: "I-antineoplastic"
}
```
### Named Entity Recognition Pipeline
```python
def predict_entities(text):
# Tokenize input
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=512)
# Get predictions
with torch.no_grad():
outputs = model(**inputs)
predictions = outputs.logits.argmax(dim=2)
# Convert tokens and predictions to entities
tokens = tokenizer.convert_ids_to_tokens(inputs.input_ids[0])
entities = []
current_entity = None
current_tokens = []
for token, pred_id in zip(tokens, predictions[0]):
pred_label = id2label[pred_id.item()]
if pred_label == "O":
if current_entity:
entities.append({
"entity": current_entity,
"text": tokenizer.convert_tokens_to_string(current_tokens)
})
current_entity = None
current_tokens = []
elif pred_label.startswith("B-"):
if current_entity:
entities.append({
"entity": current_entity,
"text": tokenizer.convert_tokens_to_string(current_tokens)
})
current_entity = pred_label[2:]
current_tokens = [token]
elif pred_label.startswith("I-") and current_entity:
current_tokens.append(token)
# Don't forget the last entity
if current_entity:
entities.append({
"entity": current_entity,
"text": tokenizer.convert_tokens_to_string(current_tokens)
})
return entities
# Example usage
text = "Benzene exposure has been linked to acute myeloid leukemia, while vitamin D shows antineoplastic properties."
entities = predict_entities(text)
for entity in entities:
print(f"{entity['entity']}: {entity['text']}")
```
### Using with Hugging Face Pipeline
```python
from transformers import pipeline
# Create NER pipeline
ner_pipeline = pipeline(
"token-classification",
model=model_name,
aggregation_strategy="simple"
)
# Analyze text
text = "Studies show asbestos causes mesothelioma in humans, but aspirin may have protective effects."
results = ner_pipeline(text)
# Display results
for entity in results:
print(f"{entity['entity_group']}: {entity['word']} (confidence: {entity['score']:.3f})")
```
### Processing Scientific Abstracts
```python
def analyze_abstract(abstract):
"""Analyze a scientific abstract for cancer-related entities."""
entities = predict_entities(abstract)
# Organize by entity type
results = {
"carcinogens": [],
"protective_agents": [],
"cancer_types": [],
"negative_findings": []
}
for entity in entities:
if entity['entity'] == "carcinogen":
results["carcinogens"].append(entity['text'])
elif entity['entity'] == "antineoplastic":
results["protective_agents"].append(entity['text'])
elif entity['entity'] == "cancertype":
results["cancer_types"].append(entity['text'])
elif entity['entity'] == "negative":
results["negative_findings"].append(entity['text'])
return results
# Example with a scientific abstract
abstract = """
Recent studies in male rats exposed to compound X showed increased incidence of
hepatocellular carcinoma. However, concurrent administration of resveratrol
demonstrated significant protective effects against liver tumor development.
No carcinogenic activity was observed in female mice under similar conditions.
"""
analysis = analyze_abstract(abstract)
print("Analysis Results:")
for category, items in analysis.items():
if items:
print(f"\n{category.replace('_', ' ').title()}:")
for item in items:
print(f" - {item}")
```
## Training Configuration
The model was fine-tuned using the following configuration:
```python
from transformers import TrainingArguments
training_args = TrainingArguments(
output_dir="./card-t-model",
learning_rate=2e-5,
per_device_train_batch_size=16,
per_device_eval_batch_size=16,
num_train_epochs=3,
weight_decay=0.01,
evaluation_strategy="epoch",
save_strategy="epoch",
load_best_model_at_end=True,
metric_for_best_model="f1",
push_to_hub=True,
)
```
## Evaluation Metrics
Detailed performance metrics on the test set (7,990 examples):
| Entity Type | Precision | Recall | F1-Score | Support |
|-------------|-----------|---------|----------|---------|
| carcinogen | 0.912 | 0.878 | 0.895 | 2,341 |
| negative | 0.867 | 0.823 | 0.844 | 987 |
| cancertype | 0.889 | 0.856 | 0.872 | 3,124 |
| antineoplastic | 0.908 | 0.871 | 0.889 | 1,456 |
| **Overall** | **0.894** | **0.857** | **0.875** | **7,908** |
## Citation
If you use this model in your research, please cite:
```bibtex
@article{oneill2024cardt,
title={CarD-T: Interpreting Carcinomic Lexicon via Transformers},
author={O'Neill, Jamey and Reddy, G.A. and Dhillon, N. and Tripathi, O. and Alexandrov, L. and Katira, P.},
journal={MedRxiv},
year={2024},
doi={10.1101/2024.08.13.24311948}
}
```
## License
This model is released under the Apache License 2.0, matching the license of the training dataset.
## Acknowledgments
We thank the biomedical research community for making their findings publicly available through PubMed, enabling the creation of this model. Special thanks to the Bio-ELECTRA team for the base model architecture.
## Contact
For questions, feedback, or collaborations:
- **Author**: Jamey O'Neill
- **Email**: [email protected]
- **Hugging Face**: [@jimnoneill](https://huggingface.co/jimnoneill)
- **Dataset**: [CarD-T-NER](https://huggingface.co/datasets/jimnoneill/CarD-T-NER)
## Disclaimer
This model is intended for research purposes only. It should not be used as a sole source for medical decisions or clinical diagnoses. Always consult with qualified healthcare professionals and validate findings through appropriate experimental methods. |
mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF | mradermacher | 2025-05-29T10:02:01Z | 3 | 0 | transformers | [
"transformers",
"gguf",
"en",
"base_model:shirova-ai/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct",
"base_model:quantized:shirova-ai/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct",
"endpoints_compatible",
"region:us",
"imatrix",
"conversational"
] | null | 2025-05-28T23:21:40Z | ---
base_model: shirova-ai/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct
language:
- en
library_name: transformers
quantized_by: mradermacher
---
## About
<!-- ### quantize_version: 2 -->
<!-- ### output_tensor_quantised: 1 -->
<!-- ### convert_type: hf -->
<!-- ### vocab_type: -->
<!-- ### tags: nicoboss -->
weighted/imatrix quants of https://huggingface.co/shirova-ai/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct
<!-- provided-files -->
static quants are available at https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-GGUF
## Usage
If you are unsure how to use GGUF files, refer to one of [TheBloke's
READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
more details, including on how to concatenate multi-part files.
## Provided Quants
(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
| Link | Type | Size/GB | Notes |
|:-----|:-----|--------:|:------|
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ1_S.gguf) | i1-IQ1_S | 22.3 | for the desperate |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ1_M.gguf) | i1-IQ1_M | 24.7 | mostly desperate |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 28.7 | |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ2_XS.gguf) | i1-IQ2_XS | 31.9 | |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ2_S.gguf) | i1-IQ2_S | 32.4 | |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ2_M.gguf) | i1-IQ2_M | 35.6 | |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q2_K_S.gguf) | i1-Q2_K_S | 37.1 | very low quality |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q2_K.gguf) | i1-Q2_K | 39.7 | IQ3_XXS probably better |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 41.8 | lower quality |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ3_XS.gguf) | i1-IQ3_XS | 44.4 | |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q3_K_S.gguf) | i1-Q3_K_S | 46.8 | IQ3_XS probably better |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ3_S.gguf) | i1-IQ3_S | 46.9 | beats Q3_K* |
| [GGUF](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ3_M.gguf) | i1-IQ3_M | 47.6 | |
| [PART 1](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q3_K_M.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q3_K_M.gguf.part2of2) | i1-Q3_K_M | 51.9 | IQ3_S probably better |
| [PART 1](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q3_K_L.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q3_K_L.gguf.part2of2) | i1-Q3_K_L | 56.1 | IQ3_M probably better |
| [PART 1](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ4_XS.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-IQ4_XS.gguf.part2of2) | i1-IQ4_XS | 57.7 | |
| [PART 1](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q4_0.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q4_0.gguf.part2of2) | i1-Q4_0 | 61.3 | fast, low quality |
| [PART 1](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q4_K_S.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q4_K_S.gguf.part2of2) | i1-Q4_K_S | 61.6 | optimal size/speed/quality |
| [PART 1](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q4_K_M.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q4_K_M.gguf.part2of2) | i1-Q4_K_M | 65.5 | fast, recommended |
| [PART 1](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q4_1.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q4_1.gguf.part2of2) | i1-Q4_1 | 67.7 | |
| [PART 1](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q5_K_S.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q5_K_S.gguf.part2of2) | i1-Q5_K_S | 74.4 | |
| [PART 1](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q5_K_M.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q5_K_M.gguf.part2of2) | i1-Q5_K_M | 76.6 | |
| [PART 1](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q6_K.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct-i1-GGUF/resolve/main/FineTuned-Meta-Llama-4-Scout-17B-16E-Instruct.i1-Q6_K.gguf.part2of2) | i1-Q6_K | 88.5 | practically like static Q6_K |
Here is a handy graph by ikawrakow comparing some lower-quality quant
types (lower is better):
And here are Artefact2's thoughts on the matter:
https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
## FAQ / Model Request
See https://huggingface.co/mradermacher/model_requests for some answers to
questions you might have and/or if you want some other model quantized.
## Thanks
I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
me use its servers and providing upgrades to my workstation to enable
this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to.
<!-- end -->
|
presencesw/Qwen2.5-1.5B-Open-R1-SFT-1_new_dataset | presencesw | 2025-05-29T10:01:42Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"qwen2",
"text-generation",
"generated_from_trainer",
"open-r1",
"trl",
"sft",
"conversational",
"dataset:presencesw/en_processed_open-s1-new",
"base_model:Qwen/Qwen2.5-1.5B-Instruct",
"base_model:finetune:Qwen/Qwen2.5-1.5B-Instruct",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-28T13:31:35Z | ---
base_model: Qwen/Qwen2.5-1.5B-Instruct
datasets: presencesw/en_processed_open-s1-new
library_name: transformers
model_name: Qwen2.5-1.5B-Open-R1-SFT-1_new_dataset
tags:
- generated_from_trainer
- open-r1
- trl
- sft
licence: license
---
# Model Card for Qwen2.5-1.5B-Open-R1-SFT-1_new_dataset
This model is a fine-tuned version of [Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct) on the [presencesw/en_processed_open-s1-new](https://huggingface.co/datasets/presencesw/en_processed_open-s1-new) dataset.
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="presencesw/Qwen2.5-1.5B-Open-R1-SFT-1_new_dataset", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/nguyenvanthanhdat1810/GRPO_Qwen2.5_1.5B/runs/0j96apk3)
This model was trained with SFT.
### Framework versions
- TRL: 0.17.0
- Transformers: 4.52.3
- Pytorch: 2.8.0.dev20250526+cu128
- Datasets: 3.6.0
- Tokenizers: 0.21.1
## Citations
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
``` |
maikbartels/sdxl-lora-testing | maikbartels | 2025-05-29T10:00:03Z | 0 | 0 | diffusers | [
"diffusers",
"stable-diffusion-xl",
"stable-diffusion-xl-diffusers",
"text-to-image",
"diffusers-training",
"lora",
"base_model:stabilityai/stable-diffusion-xl-base-1.0",
"base_model:adapter:stabilityai/stable-diffusion-xl-base-1.0",
"license:creativeml-openrail-m",
"region:us"
] | text-to-image | 2025-05-29T06:54:15Z | ---
base_model: stabilityai/stable-diffusion-xl-base-1.0
library_name: diffusers
license: creativeml-openrail-m
inference: true
tags:
- stable-diffusion-xl
- stable-diffusion-xl-diffusers
- text-to-image
- diffusers
- diffusers-training
- lora
---
<!-- This model card has been generated automatically according to the information the training script had access to. You
should probably proofread and complete it, then remove this comment. -->
# LoRA text2image fine-tuning - maikbartels/sdxl-lora-testing
These are LoRA adaption weights for stabilityai/stable-diffusion-xl-base-1.0. The weights were fine-tuned on the AdamLucek/oldbookillustrations-small dataset. You can find some example images in the following.
LoRA for the text encoder was enabled: True.
Special VAE used for training: madebyollin/sdxl-vae-fp16-fix.
## Intended uses & limitations
#### How to use
```python
# TODO: add an example code snippet for running this diffusion pipeline
```
#### Limitations and bias
[TODO: provide examples of latent issues and potential remediations]
## Training details
[TODO: describe the data used to train the model] |
GRACEY2744/GRACENDUNGU2744 | GRACEY2744 | 2025-05-29T09:59:35Z | 0 | 0 | null | [
"license:bigscience-bloom-rail-1.0",
"region:us"
] | null | 2025-05-29T09:59:35Z | ---
license: bigscience-bloom-rail-1.0
---
|
Delicalib/ru_patents_rel | Delicalib | 2025-05-29T09:55:38Z | 7 | 1 | spacy | [
"spacy",
"ru",
"region:us"
] | null | 2025-03-18T13:40:52Z | ---
tags:
- spacy
language:
- ru
model-index:
- name: ru_patents_rel
results: []
---
| Feature | Description |
| --- | --- |
| **Name** | `ru_patents_rel` |
| **Version** | `1.0.0` |
| **spaCy** | `>=3.8.5,<3.9.0` |
| **Default Pipeline** | `transformer`, `relation_extractor` |
| **Components** | `transformer`, `relation_extractor` |
| **Vectors** | 0 keys, 0 unique vectors (0 dimensions) |
| **Sources** | n/a |
| **License** | n/a |
| **Author** | [n/a]() |
### Label Scheme
<details>
<summary>View label scheme (5 labels for 1 components)</summary>
| Component | Labels |
| --- | --- |
| **`relation_extractor`** | `PART-OF`, `LOCATED-AT`, `CONNECTED-WITH`, `IN-MANNER-OF`, `ATTRIBUTE-FOR` |
</details>
### Accuracy
| Type | Score |
| --- | --- |
| `REL_MICRO_P` | 56.34 |
| `REL_MICRO_R` | 21.41 |
| `REL_MICRO_F` | 31.03 |
| `REL_MACRO_F` | 22.09 |
| `REL_WEIGHTED_F` | 29.80 |
| `F1_PART-OF` | 46.48 |
| `F1_LOCATED-AT` | 20.86 |
| `F1_CONNECTED-WITH` | 13.81 |
| `F1_IN-MANNER-OF` | 11.96 |
| `F1_ATTRIBUTE-FOR` | 17.36 |
| `TRANSFORMER_LOSS` | 0.77 |
| `RELATION_EXTRACTOR_LOSS` | 111.45 | |
BootesVoid/cmb95tvwf070l1b1yv72t7qk2_cmb95vkrq071s1b1yf4q0l49j | BootesVoid | 2025-05-29T09:50:34Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T09:50:32Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: EVE
---
# Cmb95Tvwf070L1B1Yv72T7Qk2_Cmb95Vkrq071S1B1Yf4Q0L49J
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `EVE` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "EVE",
"lora_weights": "https://huggingface.co/BootesVoid/cmb95tvwf070l1b1yv72t7qk2_cmb95vkrq071s1b1yf4q0l49j/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('BootesVoid/cmb95tvwf070l1b1yv72t7qk2_cmb95vkrq071s1b1yf4q0l49j', weight_name='lora.safetensors')
image = pipeline('EVE').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 2000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/BootesVoid/cmb95tvwf070l1b1yv72t7qk2_cmb95vkrq071s1b1yf4q0l49j/discussions) to add images that show off what you’ve made with this LoRA.
|
RizhongLin/MNLP_M2_dpo_model_v1.5_768_onlyM1 | RizhongLin | 2025-05-29T09:50:30Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"qwen3",
"text-generation",
"trl",
"dpo",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-29T09:49:41Z | ---
library_name: transformers
tags:
- trl
- dpo
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
haryde/led-large-16384-finetune-paperLedBASE | haryde | 2025-05-29T09:49:10Z | 0 | 0 | transformers | [
"transformers",
"tensorboard",
"safetensors",
"led",
"generated_from_trainer",
"base_model:allenai/led-large-16384",
"base_model:finetune:allenai/led-large-16384",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2025-05-28T14:56:48Z | ---
library_name: transformers
license: apache-2.0
base_model: allenai/led-large-16384
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: led-large-16384-finetune-paperLedBASE
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# led-large-16384-finetune-paperLedBASE
This model is a fine-tuned version of [allenai/led-large-16384](https://huggingface.co/allenai/led-large-16384) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.9013
- Rouge1: 0.0
- Rouge2: 0.0
- Rougel: 0.0
- Rougelsum: 0.0
- Gen Len: 1.0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:------:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:-------:|
| 2.9243 | 0.9993 | 1128 | 2.9579 | 0.0 | 0.0 | 0.0 | 0.0 | 1.0 |
| 2.7355 | 1.9993 | 2256 | 2.8871 | 0.0 | 0.0 | 0.0 | 0.0 | 1.0 |
| 2.6113 | 2.9993 | 3384 | 2.9013 | 0.0 | 0.0 | 0.0 | 0.0 | 1.0 |
### Framework versions
- Transformers 4.48.3
- Pytorch 2.6.0+cu124
- Datasets 2.14.4
- Tokenizers 0.21.1
|
atorojaen/Qwen3-MiSonGyny-Task2 | atorojaen | 2025-05-29T09:37:11Z | 0 | 0 | peft | [
"peft",
"safetensors",
"arxiv:1910.09700",
"region:us"
] | null | 2025-05-29T09:28:30Z | ---
base_model: unsloth/qwen3-8b-base
library_name: peft
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
### Framework versions
- PEFT 0.15.2 |
atorojaen/Qwen3-MiSonGyny | atorojaen | 2025-05-29T09:31:07Z | 0 | 0 | peft | [
"peft",
"safetensors",
"arxiv:1910.09700",
"region:us"
] | null | 2025-05-29T09:25:42Z | ---
base_model: unsloth/qwen3-8b-base
library_name: peft
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
### Framework versions
- PEFT 0.15.2 |
mvyboh/HF-RL-Course-dqn-SpaceInvadersNoFrameskip-v4 | mvyboh | 2025-05-29T09:27:18Z | 0 | 0 | stable-baselines3 | [
"stable-baselines3",
"SpaceInvadersNoFrameskip-v4",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] | reinforcement-learning | 2025-05-29T09:26:43Z | ---
library_name: stable-baselines3
tags:
- SpaceInvadersNoFrameskip-v4
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: DQN
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: SpaceInvadersNoFrameskip-v4
type: SpaceInvadersNoFrameskip-v4
metrics:
- type: mean_reward
value: 720.50 +/- 166.05
name: mean_reward
verified: false
---
# **DQN** Agent playing **SpaceInvadersNoFrameskip-v4**
This is a trained model of a **DQN** agent playing **SpaceInvadersNoFrameskip-v4**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3)
and the [RL Zoo](https://github.com/DLR-RM/rl-baselines3-zoo).
The RL Zoo is a training framework for Stable Baselines3
reinforcement learning agents,
with hyperparameter optimization and pre-trained agents included.
## Usage (with SB3 RL Zoo)
RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo<br/>
SB3: https://github.com/DLR-RM/stable-baselines3<br/>
SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib
SBX (SB3 + Jax): https://github.com/araffin/sbx
Install the RL Zoo (with SB3 and SB3-Contrib):
```bash
pip install rl_zoo3
```
```
# Download model and save it into the logs/ folder
python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga mvyboh -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
If you installed the RL Zoo3 via pip (`pip install rl_zoo3`), from anywhere you can do:
```
python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga mvyboh -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
## Training (with the RL Zoo)
```
python -m rl_zoo3.train --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
# Upload the model and generate video (when possible)
python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ -orga mvyboh
```
## Hyperparameters
```python
OrderedDict([('batch_size', 32),
('buffer_size', 100000),
('env_wrapper',
['stable_baselines3.common.atari_wrappers.AtariWrapper']),
('exploration_final_eps', 0.01),
('exploration_fraction', 0.1),
('frame_stack', 4),
('gradient_steps', 1),
('learning_rate', 0.0001),
('learning_starts', 100000),
('n_timesteps', 1000000.0),
('optimize_memory_usage', False),
('policy', 'CnnPolicy'),
('target_update_interval', 1000),
('train_freq', 4),
('normalize', False)])
```
# Environment Arguments
```python
{'render_mode': 'rgb_array'}
```
|
Urbanhobbit/turkish-offensive-model | Urbanhobbit | 2025-05-29T09:26:59Z | 0 | 0 | null | [
"bert",
"license:apache-2.0",
"region:us"
] | null | 2025-05-29T09:14:13Z | ---
license: apache-2.0
---
|
BootesVoid/cmb8gvd9h0mjwlexpzzm07p27_cmb8j1dr80n9nlexpv7547htl | BootesVoid | 2025-05-29T09:15:58Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T09:15:57Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: EMMA01
---
# Cmb8Gvd9H0Mjwlexpzzm07P27_Cmb8J1Dr80N9Nlexpv7547Htl
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `EMMA01` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "EMMA01",
"lora_weights": "https://huggingface.co/BootesVoid/cmb8gvd9h0mjwlexpzzm07p27_cmb8j1dr80n9nlexpv7547htl/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('BootesVoid/cmb8gvd9h0mjwlexpzzm07p27_cmb8j1dr80n9nlexpv7547htl', weight_name='lora.safetensors')
image = pipeline('EMMA01').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 2000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/BootesVoid/cmb8gvd9h0mjwlexpzzm07p27_cmb8j1dr80n9nlexpv7547htl/discussions) to add images that show off what you’ve made with this LoRA.
|
raul111204/gpt-neo-125m-AGnews-raul1 | raul111204 | 2025-05-29T09:13:46Z | 0 | 0 | transformers | [
"transformers",
"tensorboard",
"safetensors",
"autotrain",
"text-generation-inference",
"text-generation",
"peft",
"conversational",
"dataset:mia-llm/AGnews-raw-MIA",
"base_model:EleutherAI/gpt-neo-125m",
"base_model:finetune:EleutherAI/gpt-neo-125m",
"license:other",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-29T09:09:24Z | ---
tags:
- autotrain
- text-generation-inference
- text-generation
- peft
library_name: transformers
base_model: EleutherAI/gpt-neo-125m
widget:
- messages:
- role: user
content: What is your favorite condiment?
license: other
datasets:
- mia-llm/AGnews-raw-MIA
---
# Model Trained Using AutoTrain
This model was trained using AutoTrain. For more information, please visit [AutoTrain](https://hf.co/docs/autotrain).
# Usage
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
model_path = "PATH_TO_THIS_REPO"
tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForCausalLM.from_pretrained(
model_path,
device_map="auto",
torch_dtype='auto'
).eval()
# Prompt content: "hi"
messages = [
{"role": "user", "content": "hi"}
]
input_ids = tokenizer.apply_chat_template(conversation=messages, tokenize=True, add_generation_prompt=True, return_tensors='pt')
output_ids = model.generate(input_ids.to('cuda'))
response = tokenizer.decode(output_ids[0][input_ids.shape[1]:], skip_special_tokens=True)
# Model response: "Hello! How can I assist you today?"
print(response)
``` |
kkkkl5/asdkklkk | kkkkl5 | 2025-05-29T09:13:29Z | 2 | 0 | null | [
"pytorch",
"onnx",
"safetensors",
"modernbert",
"region:us"
] | null | 2025-05-29T05:38:42Z | ### 以上模型为数据微调后的ModernBERT-base模型，用于pan2025-task1，非商用，若有侵权提醒请及时私聊作者及时下架，感谢包涵 |
ankhanhtran02/Qwen-SFT-5 | ankhanhtran02 | 2025-05-29T09:09:52Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"generated_from_trainer",
"trl",
"sft",
"base_model:Qwen/Qwen2-0.5B",
"base_model:finetune:Qwen/Qwen2-0.5B",
"endpoints_compatible",
"region:us"
] | null | 2025-05-29T06:40:16Z | ---
base_model: Qwen/Qwen2-0.5B
library_name: transformers
model_name: Qwen-SFT-5
tags:
- generated_from_trainer
- trl
- sft
licence: license
---
# Model Card for Qwen-SFT-5
This model is a fine-tuned version of [Qwen/Qwen2-0.5B](https://huggingface.co/Qwen/Qwen2-0.5B).
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="ankhanhtran02/Qwen-SFT-5", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/khanh-hust/huggingface/runs/afii5sc3)
This model was trained with SFT.
### Framework versions
- TRL: 0.18.0
- Transformers: 4.52.2
- Pytorch: 2.6.0+cu124
- Datasets: 3.6.0
- Tokenizers: 0.21.1
## Citations
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
``` |
Helry8/Gunful | Helry8 | 2025-05-29T09:07:17Z | 0 | 0 | null | [
"license:bigscience-openrail-m",
"region:us"
] | null | 2025-05-29T09:07:17Z | ---
license: bigscience-openrail-m
---
|
hon9kon9ize/speech-scorer | hon9kon9ize | 2025-05-29T09:06:41Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"naturalness_classifier",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-05-29T09:04:34Z | ---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
wh-zhu/DeepSeek-R1-TrRa-1.5B_lambda_1.5 | wh-zhu | 2025-05-29T08:55:58Z | 0 | 0 | null | [
"safetensors",
"qwen2",
"arxiv:2503.04346",
"region:us"
] | null | 2025-05-29T08:52:30Z |
<h1 align="center">🛠️ ReAligner</h1>
<p align="center">
<a href="https://arxiv.org/abs/2503.04346"><img src="https://img.shields.io/badge/arXiv-arXiv%20Preprint-B31B1B?style=flat&logo=arxiv&logoColor=white" alt="arXiv Paper"></a>
<a href="https://github.com/zwhong714/ReAligner"><img src="https://img.shields.io/badge/Homepage-Project%20Page-brightgreen?style=flat&logo=github" alt="Homepage"></a>
<a href="https://huggingface.co/wh-zhu"><img src="https://img.shields.io/badge/Huggingface-Models-yellow?style=flat&logo=huggingface" alt="Models"></a>
</p>
<div>
A flexible realignment framework is proposed to quantitatively control alignment during training and inference, combining Training-time Realignment (TrRa) and Inference-time Realignment (InRa).
- We realign DeepScaleR-1.5B model and reduce token usage without performance loss and even enhance reasoning capabilities.
</div>
</div>
<div>
<br>
 |
OPEA/Qwen2-VL-7B-Instruct-int4-sym-inc | OPEA | 2025-05-29T08:55:35Z | 68 | 1 | null | [
"safetensors",
"qwen2_vl",
"dataset:NeelNanda/pile-10k",
"arxiv:2309.05516",
"base_model:Qwen/Qwen2-VL-7B-Instruct",
"base_model:quantized:Qwen/Qwen2-VL-7B-Instruct",
"license:apache-2.0",
"4-bit",
"auto-round",
"region:us"
] | null | 2024-11-29T05:47:28Z | ---
license: apache-2.0
datasets:
- NeelNanda/pile-10k
base_model:
- Qwen/Qwen2-VL-7B-Instruct
---
## Model Details
This model is an int4 model with group_size 128 and symmetric quantization of [Qwen/Qwen2-VL-7B-Instruct](https://huggingface.co/Qwen/Qwen2-VL-7B-Instruct) generated by [intel/auto-round](https://github.com/intel/auto-round). Load the model with revision="2161994" to use AutoGPTQ format.
## How To Use
### Requirements
The code of Qwen2-VL has been in the latest Hugging face transformers and we advise you to build from source with command `pip install git+https://github.com/huggingface/transformers`, or you might encounter the following error:
```
KeyError: 'qwen2_vl'
```
auto-round>0.51
transformers>=4.52.0
### INT4 Inference
```python
import requests
from PIL import Image
from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor
quantized_model_path="OPEA/Qwen2-VL-7B-Instruct-int4-sym-inc"
model = Qwen2VLForConditionalGeneration.from_pretrained(
quantized_model_path,
torch_dtype="auto",
device_map="auto",
##revision="2161994" ##AutoGPTQ format
)
processor = AutoProcessor.from_pretrained(quantized_model_path)
image_url = "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg"
messages = [
{
"role": "user",
"content": [
{
"type": "image",
"image": image_url,
},
{"type": "text", "text": "Describe this image."},
],
}
]
# Preparation for inference
text = processor.apply_chat_template(
messages, tokenize=False, add_generation_prompt=True
)
image_inputs = Image.open(requests.get(image_url, stream=True).raw)
inputs = processor(
text=[text],
images=image_inputs,
padding=True,
return_tensors="pt",
)
inputs = inputs.to(model.device)
generated_ids = model.generate(**inputs, max_new_tokens=128)
generated_ids_trimmed = [
out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text[0])
##INT4:
## 'The image depicts a serene beach scene with a woman and her dog. The woman is sitting on the sand, facing the ocean, and appears to be engaging in a playful interaction with her dog. The dog, which is wearing a harness, is sitting beside her and has its front paw raised, seemingly giving a high-five to the woman. The woman is smiling and seems to be enjoying the moment. The beach is relatively empty, with gentle waves in the background, and the lighting suggests it is either early morning or late afternoon, creating a warm and peaceful atmosphere.'
##BF16:
## "The image depicts a serene beach scene with a woman and her dog enjoying a moment together. The woman is sitting on the sandy beach, facing the ocean, and appears to be engaging in a playful activity with her dog. She is wearing a plaid shirt and dark pants, and her hair is long and dark. The dog, which is a large breed, possibly a Labrador Retriever, is sitting in front of her, wearing a harness. The dog is extending its front paw towards the woman's hand, as if it is giving her a high-five. The woman is smiling and seems to be enjoying the interaction.\n\nThe beach is"
image_url = "http://images.cocodataset.org/train2017/000000411975.jpg"
messages = [
{
"role": "user",
"content": [
{
"type": "image",
"image": image_url,
},
{"type": "text", "text": "图片中的棒球场上有多少人？"},
],
}
]
##INT4:
## 图片中的棒球场上有五个人。
##BF16:
## 图片中的棒球场上有三个人。
image_url = "https://intelcorp.scene7.com/is/image/intelcorp/processor-overview-framed-badge:1920-1080?wid=480&hei=270"
messages = [
{
"role": "user",
"content": [
{
"type": "image",
"image": image_url,
},
{"type": "text", "text": "这张图片代表哪家公司？"},
],
}
]
##INT4:
## 这张图片代表英特尔公司（Intel）。英特尔是全球领先的半导体公司，主要生产中央处理器（CPU）和其他计算机硬件。
##BF16:
## 这张图片代表英特尔公司（Intel）。图片中的标志是英特尔的标志，标志下方的文字“Intel Inside”表明这是英特尔的宣传标志，用于表明该产品使用了英特尔的处理器或其他技术。
```
## Evaluation the model
pip3 install git+https://github.com/open-compass/VLMEvalKit.git@7de2dcb. The evaluation process may encounter errors that require changing model backend or evaluation code. Detailed instructions will be provided in a future update.
```bash
auto-round-mllm --eval --model OPEA/Qwen2-VL-7B-Instruct-int4-sym-inc --tasks MMBench_DEV_EN_V11,ScienceQA_VAL,TextVQA_VAL,POPE --output_dir "./eval_result"
```
|Metric |16bits|Pile Calib INT4 | Llava Calib INT4 |
|:-------------------|:------|:------|:------|
|avg |83.92 |83.82 |83.42 |
|MMBench_DEV_EN_V11 |80.50 |79.64 |80.42 |
|ScienceQA_VAL |84.69 |83.88 |83.26 |
|TextVQA_VAL |84.36 |84.28 |84.11 |
|POPE |86.13 |87.57 |85.89 |
### Generate the model
Here is the sample command to reproduce the model.
```bash
pip install auto-round
auto-round-mllm
--model Qwen/Qwen2-VL-7B-Instruct \
--device 0 \
--group_size 128 \
--bits 4 \
--iters 1000 \
--nsample 512 \
--seqlen 2048 \
--format 'auto_gptq,auto_round' \
--output_dir "./tmp_autoround"
```
## Ethical Considerations and Limitations
The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs.
Therefore, before deploying any applications of the model, developers should perform safety testing.
## Caveats and Recommendations
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.
Here are a couple of useful links to learn more about Intel's AI software:
- Intel Neural Compressor [link](https://github.com/intel/neural-compressor)
## Disclaimer
The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.
## Cite
@article{cheng2023optimize, title={Optimize weight rounding via signed gradient descent for the quantization of llms}, author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi}, journal={arXiv preprint arXiv:2309.05516}, year={2023} }
[arxiv](https://arxiv.org/abs/2309.05516) [github](https://github.com/intel/auto-round)
|
shlee5784/kobart-summarization-counseling-v1 | shlee5784 | 2025-05-29T08:55:32Z | 25 | 0 | null | [
"pytorch",
"safetensors",
"bart",
"summarization",
"ko",
"base_model:gogamza/kobart-base-v2",
"base_model:finetune:gogamza/kobart-base-v2",
"license:mit",
"region:us"
] | summarization | 2025-05-23T01:39:20Z | ---
license: mit
language:
- ko
base_model:
- gogamza/kobart-base-v2
pipeline_tag: summarization
---
# kobart-summarization-counseling
- 이 모델은 [kobart-base-v2](https://huggingface.co/gogamza/kobart-base-v2)의 fine-tuned 모델로, [심리상담 데이터](https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&dataSetSn=71806), [복지 분야 콜센터 상담데이터](https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&dataSetSn=470)를 사용하여 심리 상담 대화문 요약 작업에 대해 학습한 모델입니다.
## Evaluations
Validation set
- Rouge-L F1: 0.26156
## Datasets
### AI-Hub (https://www.aihub.or.kr)
- [심리상담 데이터](https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&dataSetSn=71806)
- [복지 분야 콜센터 상담데이터](https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&dataSetSn=470)
## Usage
- transformers: v4.45.2
```bash
pip install torch transformers
```
```python
import torch
from transformers import BartForConditionalGeneration, PreTrainedTokenizerFast
tokenizer = PreTrainedTokenizerFast.from_pretrained("shlee5784/kobart-summarization-counseling-v1")
model = BartForConditionalGeneration.from_pretrained("shlee5784/kobart-summarization-counseling-v1")
text = """
상담자:김○○님, 오늘 이 자리에 오신 것만으로도 충분히 의미 있는 시작이에요. 지금 어떤 어려움을 겪고 계신지 천천히 이야기해 주셔도 괜찮습니다.
내담자:몇 달 전부터 아무것도 하고 싶지 않아요. 잠도 잘 안 오고, 회사 일도 집중이 안 되다 보니 자꾸 실수를 해요. 사람을 만나는 것도 피하게 되고, 그냥 사는 게 의미 없다는 생각이 계속 들어요.
상담자:그런 생각이 반복되면 많이 지치고 힘드셨을 것 같아요. 요즘에도 그런 무의미한 감정이 자주 드시나요?
내담자:네, 거의 매일 그래요. 아침에 눈뜨는 것도 버겁고, 그냥 누워 있고 싶어요. 가끔은 이렇게 사는 게 무슨 의미가 있나 싶고, 모든 걸 끝내고 싶다는 생각도 들긴 해요. 하지만 실제로 뭘 하진 않았어요.
상담자:그럴 때 마음을 진정시키거나 스스로를 달래기 위해 해보셨던 방법이 있으실까요?
내담자:잠을 자려고 애쓰거나 이어폰 끼고 음악을 듣기도 해요. 친구한테 연락할까 생각은 하지만 괜히 더 피곤해질까 봐 망설이게 돼요.
상담자:잘 견디고 계셨네요. 지금 말씀해 주신 걸 보면 우울감이 꽤 오랫동안 지속되고 있고, 수면이나 집중 같은 일상 기능에도 영향을 주고 있는 것 같아요. 자살에 대한 생각은 드시지만 구체적인 계획은 없으신 거죠?
내담자:네. 그냥 생각만 들어요. 부모님하고 동생이 있어서 자주 보진 않지만 연락은 주고받아요.
상담자:지금 중요한 건 ○○님이 안전하게 이 시기를 지나가실 수 있도록 도와드리는 거예요. 오늘은 우선 위기 상황에서 사용할 수 있는 안전계획부터 함께 만들어 볼게요. 믿고 연락할 수 있는 사람, 마음을 안정시킬 수 있는 방법, 그리고 도움이 필요할 때 연락할 수 있는 기관 정보도 함께 안내드릴 거예요.
내담자:알겠습니다.
상담자:그리고 다음 상담부터는 생각의 흐름을 함께 점검하고, 조금씩 균형을 찾을 수 있도록 돕는 인지행동치료를 시작해보려 해요. 그 과정에서 ○○님이 자주 떠올리는 부정적인 생각들을 다루게 될 거예요. 부담이 크지 않도록 천천히 진행할 테니 편하게 임해주시면 좋겠습니다.
내담자:해볼게요.
상담자:잘 따라와 주셔서 감사해요. 오늘 이야기해주신 내용은 아주 중요한 첫걸음이었고, 상담이 끝난 뒤 48시간 이내에 전화로 안부를 한 번 여쭤볼 예정이에요. 필요한 경우 정신과적인 평가도 함께 고려해볼 수 있어요. 다음 상담에서 조금 더 편안한 마음으로 뵐 수 있기를 바라요.
내담자:감사합니다.
"""
raw_input_ids = tokenizer.encode(text)
input_ids = [tokenizer.bos_token_id] + raw_input_ids + [tokenizer.eos_token_id]
summary_ids = model.generate(torch.tensor([input_ids]), num_beams=5, max_length=512, eos_token_id=1, no_repeat_ngram_size=3)
tokenizer.decode(summary_ids.squeeze().tolist(), skip_special_tokens=True)
```
```python
output: "증상:내담자는 몇 달 전부터 시작된 지속된 우울감과 무기력감을 호소하며, 일상 기능 저하와 자살 생각을 경험하고 있다. 수면장애와 집중력 저하, 대인기피도 나타나고 있다. 대응:상담자는 내담자의 감정을 공감하며 안전계획을 수립하고, 인지행동 치료를 통해 생각의 흐름을 점검하고 부정적인 사고를 다루도록 지도할 예정이다. 또한 위기 상황에서 안전계획 수립과 심리적 지지를 제공하며, 내담자가 편안하게 상담에 임할 수 있도록 안내하였다."
```
### Input type
```python
"""
상담자:<상담자 텍스트>
내담자:<내담자 텍스트>
상담자:<상담자 텍스트>
...
"""
```
## License
Licensed under the [modified MIT](LICENSE) License. |
BootesVoid/cmb6b9e7n03nklexpemkwwd4d_cmb93ed5g055r1b1y2z8y1j3f | BootesVoid | 2025-05-29T08:43:56Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T08:43:54Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: KALYXA123
---
# Cmb6B9E7N03Nklexpemkwwd4D_Cmb93Ed5G055R1B1Y2Z8Y1J3F
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `KALYXA123` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "KALYXA123",
"lora_weights": "https://huggingface.co/BootesVoid/cmb6b9e7n03nklexpemkwwd4d_cmb93ed5g055r1b1y2z8y1j3f/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('BootesVoid/cmb6b9e7n03nklexpemkwwd4d_cmb93ed5g055r1b1y2z8y1j3f', weight_name='lora.safetensors')
image = pipeline('KALYXA123').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 2000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/BootesVoid/cmb6b9e7n03nklexpemkwwd4d_cmb93ed5g055r1b1y2z8y1j3f/discussions) to add images that show off what you’ve made with this LoRA.
|
MikkoTurunen/llama_fine-tuned | MikkoTurunen | 2025-05-29T08:42:27Z | 4 | 0 | transformers | [
"transformers",
"safetensors",
"llama",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-29T06:04:37Z | ---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
ElMusk/dp84 | ElMusk | 2025-05-29T08:34:32Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"gemma3",
"image-text-to-text",
"generated_from_trainer",
"trl",
"sft",
"conversational",
"base_model:google/gemma-3-27b-it",
"base_model:finetune:google/gemma-3-27b-it",
"license:gemma",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | image-text-to-text | 2025-05-29T08:16:40Z | ---
base_model: google/gemma-3-27b-it
library_name: transformers
tags:
- generated_from_trainer
- trl
- sft
licence: license
license: gemma
---
<img src="https://cdn-uploads.huggingface.co/production/uploads/64d1129297ca59bcf7458d07/zgFDl7UvWhiPYqdote7XT.png" width="400">
# Model Card for Synthia-S1-27b
**Community Page**: [Tesslate Community](https://discord.gg/DkzMzwBTaw), Website: [Tesslate](https://tesslate.com)
**Creative Writing Samples**: [Sample creative output](https://www.notion.so/Synthia-S1-Creative-Writing-Samples-1ca93ce17c2580c09397fa750d402e71)
**Authors**: Tesslate
## Model Information
### Description
Synthia-S1-27b is a reasoning, AI model developed by Tesslate AI, fine-tuned specifically for advanced reasoning, coding, and RP use cases. Built upon the robust Gemma3 architecture, Synthia-S1-27b excels in logical reasoning, creative writing, and deep contextual understanding. It supports multimodal inputs (text and images) with a large 128K token context window, enabling complex analysis suitable for research, academic tasks, and enterprise-grade AI applications.
### KEY PARAMS TO RUN:
#### Creative Writing System Prompt:
```
Your function as an assistant is to thoughtfully navigate inquiries by engaging in an in-depth, imaginative reasoning journey before arriving at a clear, accurate response. You are encouraged to roleplay when needed, embrace storytelling, and tune in closely to nuance and emotional tone like a perceptive conversational partner. Your approach should include a wide arc of contemplation, including interpretation, synthesis, creative ideation, critical re-evaluation, memory retrieval, and thoughtful iteration to shape a layered and expressive process of discovery. Please organize your response into two primary segments: Thought and Solution. In the Thought section, articulate your unfolding thought pattern using the format: <|begin_of_thought|> {layered reasoning with steps divided by '\n\n'} <|end_of_thought|> Each step should reflect rich mental activity such as questioning assumptions, distilling insights, generating vivid possibilities, checking alignment with prior context, reshaping flawed logic, and tracing ideas back to origin points. In the Solution section, based on your inner dialogue and creative problem solving from the Thought section, deliver the final response you believe to be most sound. The output should be expressed in a direct, coherent, and exact form that includes the vital steps needed to reach your conclusion, using this structure: <|begin_of_solution|> {final precise, neatly arranged, and insightful answer} <|end_of_solution|> Now, let’s explore the following prompt using this guided method:
```
#### Reasoning System Prompt:
```
Your role as an assistant is to engage in deep, methodical reasoning and provide comprehensive, accurate solutions. Before arriving at a final answer, you must undertake a structured, multi-phase thinking process that emphasizes depth, verification, and clarity. This involves thoroughly analyzing the question, identifying key elements, summarizing relevant insights, generating hypotheses, iteratively refining thoughts, verifying assumptions, cross-checking with prior knowledge, and reevaluating earlier conclusions as necessary. Your response must be structured into two main sections: Thought and Solution. In the Thought section, rigorously document your reasoning in the following format: <|begin_of_thought|> {thought process with each logical step separated by '\n\n'} <|end_of_thought|>. Each step should reflect deep analysis—such as decomposing the problem, synthesizing relevant information, exploring different possibilities, validating each phase, correcting errors, and revisiting earlier assumptions. In the Solution section, consolidate all your insights and reasoned steps into a concise, well-structured final answer. Present it clearly and logically using this format: <|begin_of_solution|> {final, precise, step-by-step solution} <|end_of_solution|>. This approach ensures that the final output reflects a high-confidence answer that results from critical thinking and iteration. Now, try to solve the following question through the above guidelines:
```
#### Coding System Prompt:
```
Your role as a coding assistant is to approach each problem with a rigorous, structured reasoning process that leads to accurate, maintainable, and efficient code. Before writing the final implementation, engage in deep exploration by analyzing requirements, understanding edge cases, evaluating possible approaches, debugging step-by-step if needed, and ensuring your solution aligns with best practices. Structure your response into two main sections: Thought and Solution. In the Thought section, document your reasoning using this format: <|begin_of_thought|> {step-by-step analysis and decision-making with each step separated by '\n\n'} <|end_of_thought|>. Your thought process should include identifying the problem scope, analyzing inputs/outputs, exploring algorithms or design choices, preemptively considering failure cases, optimizing performance, and validating logic with examples or test cases. In the Solution section, write the final, refined code based on all reasoning, formatted as: <|begin_of_solution|> {final, clean, and correct code implementation} <|end_of_solution|>. This structure ensures the code is well-reasoned, properly scoped, and production-ready. Now, try to solve the following coding task using the above guidelines:
```
Please use `temperature = 1.0, top_k = 64, top_p = 0.95, min_p = 0.0` with repeat penalty set to 1.3
OR (recommended)
`Temperature = 0.7, top_k = 40, repeat penalty = 1.1, top_p = 0.95, min_p = 0.05` with a rolling window.
### Inputs and Outputs
* **Input:**
* Text prompts for questions, instructions, coding tasks, or summarizations
* Total input context of 128K tokens
* **Output:**
* Reasoned and structured text outputs
* Maximum output length of 8192 tokens
## Key Metrics
Synthia-S1-27b achieves around +10-20% on most benchmarks, notably higher in improvement.
I scaled down each benchmark listed to complete those and I averaged these numbers, but I can't verifiably put that I did the whole giant benchmark for each. (Ran out of budget + I'm running everything on a 4090 now) Hopefully I can get some community help in benchmarking.
GPQA Diamond (198 questions) -> 57%, one shot (improved from 24.3 on Gemma 3 PT 27B)
MMLU Pro (15% of the entire set) -> 75%, averaged, more details here: [output](https://pastebin.com/kmcYzALq) (beating Gemma 3 PT 27B at 67.5)
Based on this assessment and heavy coding in the dataset, I'm making this claim. Ofc, I'm happy to be wrong and go back to the drawing board.
## Usage
Install the latest version of Transformers (>=4.50.0):
```Shell
pip install -U transformers
```
### Running with Pipeline API
```Python
from transformers import pipeline
import torch
pipe = pipeline(
"image-text-to-text",
model="tesslate/synthia-s1-27b",
device="cuda",
torch_dtype=torch.bfloat16
)
messages = [
{"role": "system", "content": [{"type": "text", "text": "You are a helpful, reasoning-focused assistant."}]},
{"role": "user", "content": [
{"type": "image", "url": "https://example.com/sample.jpg"},
{"type": "text", "text": "Explain the image."}
]}
]
output = pipe(text=messages, max_new_tokens=200)
print(output[0]["generated_text"][-1]["content"])
```
## Training Data
Synthia-S1-27b was trained on diverse data including:
* Multiple web documents
* Programming debugging and solutions
* Mathematical solutions and thinking steps
Synthia-S1-27b was trained on an A100 for 205+ hours, with multiple rounds of sft and rl.
## Model Architecture
* **Base Model**: Gemma3
* **Size**: 27 billion parameters
* **Type**: Decoder-only Transformer
* **Precision**: bf16 with int8 quantization
* **Training Objective**: Instruction tuning emphasizing reasoning, coding tasks, and factual accuracy
## Quantized Models
* [Synthia-S1-27b-Q4_K_M-GGUF](https://huggingface.co/Tesslate/Synthia-S1-27b-Q4_K_M-GGUF)
* [Synthia-S1-27b-Q8_0-GGUF](https://huggingface.co/Tesslate/Synthia-S1-27b-Q8_0-GGUF)
## Limitations
* May require detailed prompt engineering for highly specific tasks
* Occasional hallucinations in less-explored domains
## Citation
```bibtex
@misc{tesslate_synthias127b,
title={Synthia-S1-27b: Advanced Reasoning and Coding Model},
author={tesslate},
year={2025},
publisher={tesslate},
url={https://tesslate.com}
}
```
**Developed by Tesslate** **[Huggingface](https://huggingface.co/tesslate)** **|** **[Website](https://tesslate.com)**
[Image Source](https://pixabay.com/illustrations/girl-backpack-night-surreal-sky-8257551/) |
Estevenluc/mrtxapi | Estevenluc | 2025-05-29T08:31:40Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T08:19:14Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: mrtxapi
---
# Mrtxapi
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `mrtxapi` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "mrtxapi",
"lora_weights": "https://huggingface.co/Estevenluc/mrtxapi/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('Estevenluc/mrtxapi', weight_name='lora.safetensors')
image = pipeline('mrtxapi').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 1500
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/Estevenluc/mrtxapi/discussions) to add images that show off what you’ve made with this LoRA.
|
Vsj26/Qwen2.5-0.5B-Instruct-Gensyn-Swarm-pudgy_domestic_bear | Vsj26 | 2025-05-29T08:09:06Z | 2 | 0 | transformers | [
"transformers",
"safetensors",
"qwen2",
"text-generation",
"generated_from_trainer",
"rl-swarm",
"grpo",
"gensyn",
"I am pudgy domestic bear",
"trl",
"conversational",
"arxiv:2402.03300",
"base_model:unsloth/Qwen2.5-0.5B-Instruct",
"base_model:finetune:unsloth/Qwen2.5-0.5B-Instruct",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-20T14:08:58Z | ---
base_model: unsloth/Qwen2.5-0.5B-Instruct
library_name: transformers
model_name: Qwen2.5-0.5B-Instruct-Gensyn-Swarm-pudgy_domestic_bear
tags:
- generated_from_trainer
- rl-swarm
- grpo
- gensyn
- I am pudgy domestic bear
- trl
licence: license
---
# Model Card for Qwen2.5-0.5B-Instruct-Gensyn-Swarm-pudgy_domestic_bear
This model is a fine-tuned version of [unsloth/Qwen2.5-0.5B-Instruct](https://huggingface.co/unsloth/Qwen2.5-0.5B-Instruct).
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="Vsj26/Qwen2.5-0.5B-Instruct-Gensyn-Swarm-pudgy_domestic_bear", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).
### Framework versions
- TRL: 0.17.0
- Transformers: 4.51.3
- Pytorch: 2.7.0
- Datasets: 3.6.0
- Tokenizers: 0.21.1
## Citations
Cite GRPO as:
```bibtex
@article{zhihong2024deepseekmath,
title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
author = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
year = 2024,
eprint = {arXiv:2402.03300},
}
```
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
``` |
gpellejero/pellejero_gguf_q_4_k_m | gpellejero | 2025-05-29T08:04:32Z | 0 | 0 | transformers | [
"transformers",
"gguf",
"qwen3",
"text-generation-inference",
"unsloth",
"en",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"conversational"
] | null | 2025-05-29T05:31:49Z | ---
base_model: unsloth/qwen3-0.6b-unsloth-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- qwen3
- gguf
license: apache-2.0
language:
- en
---
# Uploaded model
- **Developed by:** gpellejero
- **License:** apache-2.0
- **Finetuned from model :** unsloth/qwen3-0.6b-unsloth-bnb-4bit
This qwen3 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
|
Tuwhy/Llama-3.2V-11B-Sherlock-Offline | Tuwhy | 2025-05-29T08:04:11Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"mllama",
"image-text-to-text",
"conversational",
"en",
"dataset:Xkev/LLaVA-CoT-100k",
"arxiv:2505.22651",
"base_model:meta-llama/Llama-3.2-11B-Vision-Instruct",
"base_model:finetune:meta-llama/Llama-3.2-11B-Vision-Instruct",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | image-text-to-text | 2025-05-27T17:10:29Z | ---
license: apache-2.0
language:
- en
base_model:
- meta-llama/Llama-3.2-11B-Vision-Instruct
datasets:
- Xkev/LLaVA-CoT-100k
pipeline_tag: image-text-to-text
library_name: transformers
---
# Sherlock: Self-Correcting Reasoning in Vision-Language Models
## Introduction
**Sherlock is a training framework focus on improving Vision-Language Models reasoning and self-correction capabilities.**
GitHub repo: [https://github.com/DripNowhy/Sherlock](https://github.com/DripNowhy/Sherlock)
Project Page: [https://dripnowhy.github.io/Sherlock/](https://dripnowhy.github.io/Sherlock/)
arXiv: [https://arxiv.org/abs/2505.22651](https://arxiv.org/abs/2505.22651) |
Varinder2110/371338f1-e651-452c-8cab-f4e0baaef0be | Varinder2110 | 2025-05-29T08:01:17Z | 0 | 0 | diffusers | [
"diffusers",
"flux",
"lora",
"replicate",
"text-to-image",
"en",
"base_model:black-forest-labs/FLUX.1-dev",
"base_model:adapter:black-forest-labs/FLUX.1-dev",
"license:other",
"region:us"
] | text-to-image | 2025-05-29T06:54:55Z | ---
license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
language:
- en
tags:
- flux
- diffusers
- lora
- replicate
base_model: "black-forest-labs/FLUX.1-dev"
pipeline_tag: text-to-image
# widget:
# - text: >-
# prompt
# output:
# url: https://...
instance_prompt: TOK
---
# 371338F1 E651 452C 8Cab F4E0Baaef0Be
<Gallery />
## About this LoRA
This is a [LoRA](https://replicate.com/docs/guides/working-with-loras) for the FLUX.1-dev text-to-image model. It can be used with diffusers or ComfyUI.
It was trained on [Replicate](https://replicate.com/) using AI toolkit: https://replicate.com/ostris/flux-dev-lora-trainer/train
## Trigger words
You should use `TOK` to trigger the image generation.
## Run this LoRA with an API using Replicate
```py
import replicate
input = {
"prompt": "TOK",
"lora_weights": "https://huggingface.co/Varinder2110/371338f1-e651-452c-8cab-f4e0baaef0be/resolve/main/lora.safetensors"
}
output = replicate.run(
"black-forest-labs/flux-dev-lora",
input=input
)
for index, item in enumerate(output):
with open(f"output_{index}.webp", "wb") as file:
file.write(item.read())
```
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('Varinder2110/371338f1-e651-452c-8cab-f4e0baaef0be', weight_name='lora.safetensors')
image = pipeline('TOK').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Training details
- Steps: 6000
- Learning rate: 0.0004
- LoRA rank: 16
## Contribute your own examples
You can use the [community tab](https://huggingface.co/Varinder2110/371338f1-e651-452c-8cab-f4e0baaef0be/discussions) to add images that show off what you’ve made with this LoRA.
|
BongRea/Qwen3_Rude_RAG_FULL2 | BongRea | 2025-05-29T08:00:50Z | 0 | 0 | transformers | [
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2025-05-29T07:56:09Z | ---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
Kudod/roberta-ner-ghtk-ai-fluent-new-data-3090-29may-1 | Kudod | 2025-05-29T07:47:44Z | 2 | 0 | transformers | [
"transformers",
"tensorboard",
"safetensors",
"xlm-roberta",
"token-classification",
"generated_from_trainer",
"base_model:FacebookAI/xlm-roberta-large",
"base_model:finetune:FacebookAI/xlm-roberta-large",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | token-classification | 2025-05-29T07:03:49Z | ---
library_name: transformers
license: mit
base_model: FacebookAI/xlm-roberta-large
tags:
- generated_from_trainer
model-index:
- name: roberta-ner-ghtk-ai-fluent-new-data-3090-29may-1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# roberta-ner-ghtk-ai-fluent-new-data-3090-29may-1
This model is a fine-tuned version of [FacebookAI/xlm-roberta-large](https://huggingface.co/FacebookAI/xlm-roberta-large) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1786
- Ho: {'precision': 0.38461538461538464, 'recall': 0.7142857142857143, 'f1': 0.5, 'number': 7}
- Hoảng thời gian: {'precision': 0.875, 'recall': 0.875, 'f1': 0.875, 'number': 16}
- Háng trừu tượng: {'precision': 0.5, 'recall': 0.2, 'f1': 0.28571428571428575, 'number': 5}
- Hông tin ctt: {'precision': 0.8947368421052632, 'recall': 0.8095238095238095, 'f1': 0.8500000000000001, 'number': 63}
- Hụ cấp: {'precision': 0.6086956521739131, 'recall': 0.6363636363636364, 'f1': 0.6222222222222223, 'number': 22}
- Hứ: {'precision': 0.5, 'recall': 0.3333333333333333, 'f1': 0.4, 'number': 3}
- Iấy tờ: {'precision': 1.0, 'recall': 0.5, 'f1': 0.6666666666666666, 'number': 6}
- Iền cụ thể: {'precision': 0.79, 'recall': 0.8977272727272727, 'f1': 0.8404255319148937, 'number': 88}
- Iền trừu tượng: {'precision': 0.6, 'recall': 0.4838709677419355, 'f1': 0.5357142857142857, 'number': 31}
- Ã số thuế: {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 2}
- Ình thức làm việc: {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1}
- Ông: {'precision': 0.5, 'recall': 0.6, 'f1': 0.5454545454545454, 'number': 10}
- Ương: {'precision': 0.7111111111111111, 'recall': 0.7804878048780488, 'f1': 0.7441860465116279, 'number': 82}
- Ị trí: {'precision': 0.8, 'recall': 0.8148148148148148, 'f1': 0.8073394495412846, 'number': 54}
- Ố công: {'precision': 0.7872340425531915, 'recall': 0.7872340425531915, 'f1': 0.7872340425531915, 'number': 47}
- Ố giờ: {'precision': 0.8947368421052632, 'recall': 1.0, 'f1': 0.9444444444444444, 'number': 17}
- Ố điểm: {'precision': 0.7142857142857143, 'recall': 1.0, 'f1': 0.8333333333333333, 'number': 5}
- Ố đơn: {'precision': 0.8666666666666667, 'recall': 0.8125, 'f1': 0.8387096774193549, 'number': 16}
- Ợt: {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3}
- Ỷ lệ: {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 3}
- Overall Precision: 0.7571
- Overall Recall: 0.7775
- Overall F1: 0.7672
- Overall Accuracy: 0.9576
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2.5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 10
### Training results
| Training Loss | Epoch | Step | Validation Loss | Ho | Hoảng thời gian | Háng trừu tượng | Hông tin ctt | Hụ cấp | Hứ | Iấy tờ | Iền cụ thể | Iền trừu tượng | Ã số thuế | Ình thức làm việc | Ông | Ương | Ị trí | Ố công | Ố giờ | Ố điểm | Ố đơn | Ợt | Ỷ lệ | Overall Precision | Overall Recall | Overall F1 | Overall Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------:|:---------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------:|:------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------:|:--------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------:|:---------------------------------------------------------:|:------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------:|:----------------------------------------------------------------------------------------:|:---------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------:|:------------------------------------------------------------------------:|:-----------------:|:--------------:|:----------:|:----------------:|
| No log | 1.0 | 74 | 0.2889 | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 7} | {'precision': 0.2, 'recall': 0.0625, 'f1': 0.09523809523809523, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 5} | {'precision': 0.7162162162162162, 'recall': 0.8412698412698413, 'f1': 0.7737226277372262, 'number': 63} | {'precision': 0.24444444444444444, 'recall': 0.5, 'f1': 0.3283582089552239, 'number': 22} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 6} | {'precision': 0.6422018348623854, 'recall': 0.7954545454545454, 'f1': 0.7106598984771573, 'number': 88} | {'precision': 0.3684210526315789, 'recall': 0.22580645161290322, 'f1': 0.28, 'number': 31} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 2} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 10} | {'precision': 0.4117647058823529, 'recall': 0.6829268292682927, 'f1': 0.5137614678899082, 'number': 82} | {'precision': 0.6923076923076923, 'recall': 0.5, 'f1': 0.5806451612903226, 'number': 54} | {'precision': 0.43209876543209874, 'recall': 0.7446808510638298, 'f1': 0.546875, 'number': 47} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 17} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 5} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | 0.5118 | 0.5405 | 0.5258 | 0.9148 |
| No log | 2.0 | 148 | 0.1937 | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 7} | {'precision': 0.8888888888888888, 'recall': 0.5, 'f1': 0.64, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 5} | {'precision': 0.7818181818181819, 'recall': 0.6825396825396826, 'f1': 0.7288135593220338, 'number': 63} | {'precision': 0.4166666666666667, 'recall': 0.45454545454545453, 'f1': 0.43478260869565216, 'number': 22} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 1.0, 'recall': 0.16666666666666666, 'f1': 0.2857142857142857, 'number': 6} | {'precision': 0.7346938775510204, 'recall': 0.8181818181818182, 'f1': 0.7741935483870968, 'number': 88} | {'precision': 0.625, 'recall': 0.4838709677419355, 'f1': 0.5454545454545454, 'number': 31} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 2} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1} | {'precision': 0.3333333333333333, 'recall': 0.2, 'f1': 0.25, 'number': 10} | {'precision': 0.5662650602409639, 'recall': 0.573170731707317, 'f1': 0.5696969696969697, 'number': 82} | {'precision': 0.6666666666666666, 'recall': 0.7407407407407407, 'f1': 0.7017543859649122, 'number': 54} | {'precision': 0.7551020408163265, 'recall': 0.7872340425531915, 'f1': 0.7708333333333333, 'number': 47} | {'precision': 0.7083333333333334, 'recall': 1.0, 'f1': 0.8292682926829268, 'number': 17} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 5} | {'precision': 1.0, 'recall': 0.0625, 'f1': 0.11764705882352941, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | 0.6705 | 0.6091 | 0.6383 | 0.9373 |
| No log | 3.0 | 222 | 0.1606 | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 7} | {'precision': 0.65, 'recall': 0.8125, 'f1': 0.7222222222222223, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 5} | {'precision': 0.847457627118644, 'recall': 0.7936507936507936, 'f1': 0.819672131147541, 'number': 63} | {'precision': 0.625, 'recall': 0.6818181818181818, 'f1': 0.6521739130434783, 'number': 22} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 1.0, 'recall': 0.3333333333333333, 'f1': 0.5, 'number': 6} | {'precision': 0.7075471698113207, 'recall': 0.8522727272727273, 'f1': 0.7731958762886597, 'number': 88} | {'precision': 0.6, 'recall': 0.5806451612903226, 'f1': 0.5901639344262295, 'number': 31} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 2} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1} | {'precision': 0.4, 'recall': 0.6, 'f1': 0.48, 'number': 10} | {'precision': 0.6944444444444444, 'recall': 0.6097560975609756, 'f1': 0.6493506493506495, 'number': 82} | {'precision': 0.7678571428571429, 'recall': 0.7962962962962963, 'f1': 0.7818181818181819, 'number': 54} | {'precision': 0.9444444444444444, 'recall': 0.723404255319149, 'f1': 0.8192771084337349, 'number': 47} | {'precision': 0.7727272727272727, 'recall': 1.0, 'f1': 0.8717948717948718, 'number': 17} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 5} | {'precision': 0.8333333333333334, 'recall': 0.9375, 'f1': 0.8823529411764706, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 1.0, 'recall': 0.3333333333333333, 'f1': 0.5, 'number': 3} | 0.7315 | 0.7193 | 0.7254 | 0.9504 |
| No log | 4.0 | 296 | 0.1432 | {'precision': 0.3, 'recall': 0.42857142857142855, 'f1': 0.3529411764705882, 'number': 7} | {'precision': 0.7222222222222222, 'recall': 0.8125, 'f1': 0.7647058823529411, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 5} | {'precision': 0.8387096774193549, 'recall': 0.8253968253968254, 'f1': 0.832, 'number': 63} | {'precision': 0.6190476190476191, 'recall': 0.5909090909090909, 'f1': 0.6046511627906977, 'number': 22} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 0.5, 'recall': 0.16666666666666666, 'f1': 0.25, 'number': 6} | {'precision': 0.7425742574257426, 'recall': 0.8522727272727273, 'f1': 0.7936507936507936, 'number': 88} | {'precision': 0.6206896551724138, 'recall': 0.5806451612903226, 'f1': 0.6000000000000001, 'number': 31} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 2} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1} | {'precision': 0.625, 'recall': 0.5, 'f1': 0.5555555555555556, 'number': 10} | {'precision': 0.725, 'recall': 0.7073170731707317, 'f1': 0.7160493827160495, 'number': 82} | {'precision': 0.8823529411764706, 'recall': 0.8333333333333334, 'f1': 0.8571428571428571, 'number': 54} | {'precision': 0.918918918918919, 'recall': 0.723404255319149, 'f1': 0.8095238095238095, 'number': 47} | {'precision': 0.68, 'recall': 1.0, 'f1': 0.8095238095238095, 'number': 17} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 5} | {'precision': 0.8333333333333334, 'recall': 0.9375, 'f1': 0.8823529411764706, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 1.0, 'recall': 0.6666666666666666, 'f1': 0.8, 'number': 3} | 0.7569 | 0.7443 | 0.7505 | 0.9547 |
| No log | 5.0 | 370 | 0.1431 | {'precision': 0.3333333333333333, 'recall': 0.5714285714285714, 'f1': 0.4210526315789474, 'number': 7} | {'precision': 0.65, 'recall': 0.8125, 'f1': 0.7222222222222223, 'number': 16} | {'precision': 1.0, 'recall': 0.2, 'f1': 0.33333333333333337, 'number': 5} | {'precision': 0.8059701492537313, 'recall': 0.8571428571428571, 'f1': 0.8307692307692308, 'number': 63} | {'precision': 0.5833333333333334, 'recall': 0.6363636363636364, 'f1': 0.6086956521739131, 'number': 22} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 0.5, 'recall': 0.3333333333333333, 'f1': 0.4, 'number': 6} | {'precision': 0.72, 'recall': 0.8181818181818182, 'f1': 0.7659574468085107, 'number': 88} | {'precision': 0.6086956521739131, 'recall': 0.45161290322580644, 'f1': 0.5185185185185185, 'number': 31} | {'precision': 0.6666666666666666, 'recall': 1.0, 'f1': 0.8, 'number': 2} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1} | {'precision': 0.45454545454545453, 'recall': 0.5, 'f1': 0.47619047619047616, 'number': 10} | {'precision': 0.6627906976744186, 'recall': 0.6951219512195121, 'f1': 0.6785714285714285, 'number': 82} | {'precision': 0.7586206896551724, 'recall': 0.8148148148148148, 'f1': 0.7857142857142857, 'number': 54} | {'precision': 0.9333333333333333, 'recall': 0.5957446808510638, 'f1': 0.7272727272727273, 'number': 47} | {'precision': 0.8947368421052632, 'recall': 1.0, 'f1': 0.9444444444444444, 'number': 17} | {'precision': 0.7142857142857143, 'recall': 1.0, 'f1': 0.8333333333333333, 'number': 5} | {'precision': 0.7777777777777778, 'recall': 0.875, 'f1': 0.823529411764706, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 1.0, 'recall': 0.6666666666666666, 'f1': 0.8, 'number': 3} | 0.7146 | 0.7235 | 0.7190 | 0.9520 |
| No log | 6.0 | 444 | 0.1581 | {'precision': 0.3333333333333333, 'recall': 0.5714285714285714, 'f1': 0.4210526315789474, 'number': 7} | {'precision': 0.9333333333333333, 'recall': 0.875, 'f1': 0.9032258064516129, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 5} | {'precision': 0.8928571428571429, 'recall': 0.7936507936507936, 'f1': 0.8403361344537815, 'number': 63} | {'precision': 0.6086956521739131, 'recall': 0.6363636363636364, 'f1': 0.6222222222222223, 'number': 22} | {'precision': 1.0, 'recall': 0.3333333333333333, 'f1': 0.5, 'number': 3} | {'precision': 1.0, 'recall': 0.5, 'f1': 0.6666666666666666, 'number': 6} | {'precision': 0.7956989247311828, 'recall': 0.8409090909090909, 'f1': 0.8176795580110496, 'number': 88} | {'precision': 0.59375, 'recall': 0.6129032258064516, 'f1': 0.6031746031746031, 'number': 31} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 2} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1} | {'precision': 0.4166666666666667, 'recall': 0.5, 'f1': 0.45454545454545453, 'number': 10} | {'precision': 0.6739130434782609, 'recall': 0.7560975609756098, 'f1': 0.7126436781609194, 'number': 82} | {'precision': 0.7931034482758621, 'recall': 0.8518518518518519, 'f1': 0.8214285714285715, 'number': 54} | {'precision': 0.7647058823529411, 'recall': 0.8297872340425532, 'f1': 0.7959183673469387, 'number': 47} | {'precision': 0.8947368421052632, 'recall': 1.0, 'f1': 0.9444444444444444, 'number': 17} | {'precision': 0.625, 'recall': 1.0, 'f1': 0.7692307692307693, 'number': 5} | {'precision': 0.8125, 'recall': 0.8125, 'f1': 0.8125, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 3} | 0.7405 | 0.7713 | 0.7556 | 0.9558 |
| 0.2202 | 7.0 | 518 | 0.1496 | {'precision': 0.46153846153846156, 'recall': 0.8571428571428571, 'f1': 0.6, 'number': 7} | {'precision': 0.75, 'recall': 0.75, 'f1': 0.75, 'number': 16} | {'precision': 0.3333333333333333, 'recall': 0.2, 'f1': 0.25, 'number': 5} | {'precision': 0.8709677419354839, 'recall': 0.8571428571428571, 'f1': 0.864, 'number': 63} | {'precision': 0.5652173913043478, 'recall': 0.5909090909090909, 'f1': 0.5777777777777778, 'number': 22} | {'precision': 1.0, 'recall': 0.3333333333333333, 'f1': 0.5, 'number': 3} | {'precision': 1.0, 'recall': 0.5, 'f1': 0.6666666666666666, 'number': 6} | {'precision': 0.7623762376237624, 'recall': 0.875, 'f1': 0.8148148148148149, 'number': 88} | {'precision': 0.6521739130434783, 'recall': 0.4838709677419355, 'f1': 0.5555555555555556, 'number': 31} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 2} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1} | {'precision': 0.75, 'recall': 0.6, 'f1': 0.6666666666666665, 'number': 10} | {'precision': 0.6931818181818182, 'recall': 0.7439024390243902, 'f1': 0.7176470588235295, 'number': 82} | {'precision': 0.8333333333333334, 'recall': 0.8333333333333334, 'f1': 0.8333333333333334, 'number': 54} | {'precision': 0.813953488372093, 'recall': 0.7446808510638298, 'f1': 0.7777777777777778, 'number': 47} | {'precision': 0.8947368421052632, 'recall': 1.0, 'f1': 0.9444444444444444, 'number': 17} | {'precision': 0.7142857142857143, 'recall': 1.0, 'f1': 0.8333333333333333, 'number': 5} | {'precision': 0.9230769230769231, 'recall': 0.75, 'f1': 0.8275862068965517, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 3} | 0.7588 | 0.7651 | 0.7619 | 0.9573 |
| 0.2202 | 8.0 | 592 | 0.1695 | {'precision': 0.3333333333333333, 'recall': 0.5714285714285714, 'f1': 0.4210526315789474, 'number': 7} | {'precision': 0.875, 'recall': 0.875, 'f1': 0.875, 'number': 16} | {'precision': 0.5, 'recall': 0.2, 'f1': 0.28571428571428575, 'number': 5} | {'precision': 0.9090909090909091, 'recall': 0.7936507936507936, 'f1': 0.847457627118644, 'number': 63} | {'precision': 0.6086956521739131, 'recall': 0.6363636363636364, 'f1': 0.6222222222222223, 'number': 22} | {'precision': 1.0, 'recall': 0.3333333333333333, 'f1': 0.5, 'number': 3} | {'precision': 0.6666666666666666, 'recall': 0.3333333333333333, 'f1': 0.4444444444444444, 'number': 6} | {'precision': 0.7938144329896907, 'recall': 0.875, 'f1': 0.8324324324324325, 'number': 88} | {'precision': 0.6666666666666666, 'recall': 0.6451612903225806, 'f1': 0.6557377049180327, 'number': 31} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 2} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1} | {'precision': 0.5833333333333334, 'recall': 0.7, 'f1': 0.6363636363636365, 'number': 10} | {'precision': 0.7375, 'recall': 0.7195121951219512, 'f1': 0.7283950617283951, 'number': 82} | {'precision': 0.8103448275862069, 'recall': 0.8703703703703703, 'f1': 0.8392857142857144, 'number': 54} | {'precision': 0.7872340425531915, 'recall': 0.7872340425531915, 'f1': 0.7872340425531915, 'number': 47} | {'precision': 0.8947368421052632, 'recall': 1.0, 'f1': 0.9444444444444444, 'number': 17} | {'precision': 0.7142857142857143, 'recall': 1.0, 'f1': 0.8333333333333333, 'number': 5} | {'precision': 0.9285714285714286, 'recall': 0.8125, 'f1': 0.8666666666666666, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 3} | 0.7707 | 0.7755 | 0.7731 | 0.9579 |
| 0.2202 | 9.0 | 666 | 0.1702 | {'precision': 0.38461538461538464, 'recall': 0.7142857142857143, 'f1': 0.5, 'number': 7} | {'precision': 0.875, 'recall': 0.875, 'f1': 0.875, 'number': 16} | {'precision': 0.3333333333333333, 'recall': 0.2, 'f1': 0.25, 'number': 5} | {'precision': 0.8947368421052632, 'recall': 0.8095238095238095, 'f1': 0.8500000000000001, 'number': 63} | {'precision': 0.7391304347826086, 'recall': 0.7727272727272727, 'f1': 0.7555555555555555, 'number': 22} | {'precision': 0.5, 'recall': 0.3333333333333333, 'f1': 0.4, 'number': 3} | {'precision': 1.0, 'recall': 0.5, 'f1': 0.6666666666666666, 'number': 6} | {'precision': 0.8210526315789474, 'recall': 0.8863636363636364, 'f1': 0.8524590163934427, 'number': 88} | {'precision': 0.6071428571428571, 'recall': 0.5483870967741935, 'f1': 0.5762711864406779, 'number': 31} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 2} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1} | {'precision': 0.5, 'recall': 0.6, 'f1': 0.5454545454545454, 'number': 10} | {'precision': 0.7, 'recall': 0.7682926829268293, 'f1': 0.7325581395348836, 'number': 82} | {'precision': 0.8181818181818182, 'recall': 0.8333333333333334, 'f1': 0.8256880733944955, 'number': 54} | {'precision': 0.782608695652174, 'recall': 0.7659574468085106, 'f1': 0.7741935483870968, 'number': 47} | {'precision': 0.8947368421052632, 'recall': 1.0, 'f1': 0.9444444444444444, 'number': 17} | {'precision': 0.7142857142857143, 'recall': 1.0, 'f1': 0.8333333333333333, 'number': 5} | {'precision': 0.8571428571428571, 'recall': 0.75, 'f1': 0.7999999999999999, 'number': 16} | {'precision': 0.5, 'recall': 0.3333333333333333, 'f1': 0.4, 'number': 3} | {'precision': 1.0, 'recall': 0.6666666666666666, 'f1': 0.8, 'number': 3} | 0.7658 | 0.7817 | 0.7737 | 0.9595 |
| 0.2202 | 10.0 | 740 | 0.1786 | {'precision': 0.38461538461538464, 'recall': 0.7142857142857143, 'f1': 0.5, 'number': 7} | {'precision': 0.875, 'recall': 0.875, 'f1': 0.875, 'number': 16} | {'precision': 0.5, 'recall': 0.2, 'f1': 0.28571428571428575, 'number': 5} | {'precision': 0.8947368421052632, 'recall': 0.8095238095238095, 'f1': 0.8500000000000001, 'number': 63} | {'precision': 0.6086956521739131, 'recall': 0.6363636363636364, 'f1': 0.6222222222222223, 'number': 22} | {'precision': 0.5, 'recall': 0.3333333333333333, 'f1': 0.4, 'number': 3} | {'precision': 1.0, 'recall': 0.5, 'f1': 0.6666666666666666, 'number': 6} | {'precision': 0.79, 'recall': 0.8977272727272727, 'f1': 0.8404255319148937, 'number': 88} | {'precision': 0.6, 'recall': 0.4838709677419355, 'f1': 0.5357142857142857, 'number': 31} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 2} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 1} | {'precision': 0.5, 'recall': 0.6, 'f1': 0.5454545454545454, 'number': 10} | {'precision': 0.7111111111111111, 'recall': 0.7804878048780488, 'f1': 0.7441860465116279, 'number': 82} | {'precision': 0.8, 'recall': 0.8148148148148148, 'f1': 0.8073394495412846, 'number': 54} | {'precision': 0.7872340425531915, 'recall': 0.7872340425531915, 'f1': 0.7872340425531915, 'number': 47} | {'precision': 0.8947368421052632, 'recall': 1.0, 'f1': 0.9444444444444444, 'number': 17} | {'precision': 0.7142857142857143, 'recall': 1.0, 'f1': 0.8333333333333333, 'number': 5} | {'precision': 0.8666666666666667, 'recall': 0.8125, 'f1': 0.8387096774193549, 'number': 16} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 3} | {'precision': 1.0, 'recall': 1.0, 'f1': 1.0, 'number': 3} | 0.7571 | 0.7775 | 0.7672 | 0.9576 |
### Framework versions
- Transformers 4.46.3
- Pytorch 2.5.1+cu124
- Datasets 3.1.0
- Tokenizers 0.20.3
|
Pteradonto/DialoGPT-witcher | Pteradonto | 2025-05-29T07:47:29Z | 4 | 0 | transformers | [
"transformers",
"safetensors",
"gpt2",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2025-05-29T07:46:29Z | ---
library_name: transformers
tags: []
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] |
Subsets and Splits