modelId
stringlengths 5
139
| author
stringlengths 2
42
| last_modified
timestamp[us, tz=UTC]date 2020-02-15 11:33:14
2025-06-02 12:28:20
| downloads
int64 0
223M
| likes
int64 0
11.7k
| library_name
stringclasses 462
values | tags
sequencelengths 1
4.05k
| pipeline_tag
stringclasses 54
values | createdAt
timestamp[us, tz=UTC]date 2022-03-02 23:29:04
2025-06-02 12:26:48
| card
stringlengths 11
1.01M
|
|---|---|---|---|---|---|---|---|---|---|
jason1i/speecht5_finetuned_voxpopuli_nl | jason1i | 2023-07-09T21:41:07Z | 76 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"speecht5",
"text-to-audio",
"generated_from_trainer",
"nl",
"dataset:facebook/voxpopuli",
"license:mit",
"endpoints_compatible",
"region:us"
] | text-to-audio | 2023-07-09T20:40:39Z | ---
language:
- nl
license: mit
tags:
- generated_from_trainer
datasets:
- facebook/voxpopuli
model-index:
- name: speecht5_finetuned_voxpopuli_nl
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. -->
# speecht5_finetuned_voxpopuli_nl
This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the voxpopuli dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4597
## 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: 4
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 0.5265 | 4.3 | 1000 | 0.4792 |
| 0.5034 | 8.61 | 2000 | 0.4672 |
| 0.4932 | 12.91 | 3000 | 0.4616 |
| 0.4882 | 17.21 | 4000 | 0.4597 |
### Framework versions
- Transformers 4.30.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
pavankantharaju/ppo-LunarLander-v2 | pavankantharaju | 2023-07-09T21:39:58Z | 2 | 0 | stable-baselines3 | [
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] | reinforcement-learning | 2023-07-09T21:39:40Z | ---
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: 269.66 +/- 17.48
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
...
```
|
FFusion/FFUSION.ai-Text-Encoder-LyCORIS-SD-2.1 | FFusion | 2023-07-09T21:22:09Z | 0 | 1 | null | [
"di.ffusion.ai",
"stable-diffusion",
"LyCORIS",
"LoRA",
"en",
"arxiv:1910.09700",
"arxiv:2108.06098",
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-06-06T18:21:28Z | ---
license: creativeml-openrail-m
language:
- en
tags:
- di.ffusion.ai
- stable-diffusion
- LyCORIS
- LoRA
---
# Model Card for di.FFUSION.ai Text Encoder - SD 2.1 LyCORIS
<!-- Provide a quick summary of what the model is/does. [Optional] -->
di.FFUSION.ai-tXe-FXAA
Trained on "121361" images.
- **DOWNLOAD:** https://huggingface.co/FFusion/FFUSION.ai-Text-Encoder-LyCORIS-SD-2.1/blob/main/di.FFUSION.ai-tXe-FXAA.safetensors
Enhance your model's quality and sharpness using your own pre-trained Unet.
The text encoder (without UNET) is wrapped in LyCORIS. Optimizer: torch.optim.adamw.AdamW(weight_decay=0.01, betas=(0.9, 0.99))
Network dimension/rank: 768.0 Alpha: 768.0 Module: lycoris.kohya {'conv_dim': '256', 'conv_alpha': '256', 'algo': 'loha'}
Large size due to Lyco CONV 256
This is a heavy experimental version we used to test even with sloppy captions (quick WD tags and terrible clip), yet the results were satisfying.
Note: This is not the text encoder used in the official FFUSION AI model.
# SAMPLES
**Available also at https://civitai.com/models/83622**
For a1111
Install https://github.com/KohakuBlueleaf/a1111-sd-webui-lycoris
Download di.FFUSION.ai-tXe-FXAA to /models/Lycoris
Option1:
Insert <lyco:di.FFUSION.ai-tXe-FXAA:1.0> to prompt
No need to split Unet and Text Enc as its only TX encoder there.
You can go up to 2x weights
Option2: If you need it always ON (ex run a batch from txt file) then you can go to settings / Quicksettings list
add sd_lyco
restart and you should have a drop-down now 🤟 🥃
# Table of Contents
- [Model Card for di.FFUSION.ai Text Encoder - SD 2.1 LyCORIS](#model-card-for--model_id-)
- [Table of Contents](#table-of-contents)
- [Table of Contents](#table-of-contents-1)
- [Model Details](#model-details)
- [Model Description](#model-description)
- [Uses](#uses)
- [Direct Use](#direct-use)
- [Downstream Use [Optional]](#downstream-use-optional)
- [Out-of-Scope Use](#out-of-scope-use)
- [Bias, Risks, and Limitations](#bias-risks-and-limitations)
- [Recommendations](#recommendations)
- [Training Details](#training-details)
- [Training Data](#training-data)
- [Training Procedure](#training-procedure)
- [Preprocessing](#preprocessing)
- [Speeds, Sizes, Times](#speeds-sizes-times)
- [Evaluation](#evaluation)
- [Testing Data, Factors & Metrics](#testing-data-factors--metrics)
- [Testing Data](#testing-data)
- [Factors](#factors)
- [Metrics](#metrics)
- [Results](#results)
- [Model Examination](#model-examination)
- [Environmental Impact](#environmental-impact)
- [Technical Specifications [optional]](#technical-specifications-optional)
- [Model Architecture and Objective](#model-architecture-and-objective)
- [Compute Infrastructure](#compute-infrastructure)
- [Hardware](#hardware)
- [Software](#software)
- [Citation](#citation)
- [Glossary [optional]](#glossary-optional)
- [More Information [optional]](#more-information-optional)
- [Model Card Authors [optional]](#model-card-authors-optional)
- [Model Card Contact](#model-card-contact)
- [How to Get Started with the Model](#how-to-get-started-with-the-model)
# Model Details
## Model Description
<!-- Provide a longer summary of what this model is/does. -->
di.FFUSION.ai-tXe-FXAA
Trained on "121361" images.
Enhance your model's quality and sharpness using your own pre-trained Unet.
The text encoder (without UNET) is wrapped in LyCORIS. Optimizer: torch.optim.adamw.AdamW(weight_decay=0.01, betas=(0.9, 0.99))
Network dimension/rank: 768.0 Alpha: 768.0 Module: lycoris.kohya {'conv_dim': '256', 'conv_alpha': '256', 'algo': 'loha'}
Large size due to Lyco CONV 256
This is a heavy experimental version we used to test even with sloppy captions (quick WD tags and terrible clip), yet the results were satisfying.
Note: This is not the text encoder used in the official FFUSION AI model.
- **Developed by:** FFusion.ai
- **Shared by [Optional]:** idle stoev
- **Model type:** Language model
- **Language(s) (NLP):** en
- **License:** creativeml-openrail-m
- **Parent Model:** More information needed
- **Resources for more information:** 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. -->
<!-- If the user enters content, print that. If not, but they enter a task in the list, use that. If neither, say "more info needed." -->
The text encoder (without UNET) is wrapped in LyCORIS. Optimizer: torch.optim.adamw.AdamW(weight_decay=0.01, betas=(0.9, 0.99))
Network dimension/rank: 768.0 Alpha: 768.0 Module: lycoris.kohya {'conv_dim': '256', 'conv_alpha': '256', 'algo': 'loha'}
Large size due to Lyco CONV 256
# Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)). Predictions generated by the model may include disturbing and harmful stereotypes across protected classes; identity characteristics; and sensitive, social, and occupational groups.
## Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
# Training Details
## Training Data
<!-- This should link to a Data 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. -->
Trained on "121361" images.
ss_caption_tag_dropout_rate: "0.0",
ss_multires_noise_discount: "0.3",
ss_mixed_precision: "bf16",
ss_text_encoder_lr: "1e-07",
ss_keep_tokens: "3",
ss_network_args: "{"conv_dim": "256", "conv_alpha": "256", "algo": "loha"}",
ss_caption_dropout_rate: "0.02",
ss_flip_aug: "False",
ss_learning_rate: "2e-07",
ss_sd_model_name: "stabilityai/stable-diffusion-2-1-base",
ss_max_grad_norm: "1.0",
ss_num_epochs: "2",
ss_gradient_checkpointing: "False",
ss_face_crop_aug_range: "None",
ss_epoch: "2",
ss_num_train_images: "121361",
ss_color_aug: "False",
ss_gradient_accumulation_steps: "1",
ss_total_batch_size: "100",
ss_prior_loss_weight: "1.0",
ss_training_comment: "None",
ss_network_dim: "768",
ss_output_name: "FusionaMEGA1tX",
ss_max_bucket_reso: "1024",
ss_network_alpha: "768.0",
ss_steps: "2444",
ss_shuffle_caption: "True",
ss_training_finished_at: "1684158038.0763328",
ss_min_bucket_reso: "256",
ss_noise_offset: "0.09",
ss_enable_bucket: "True",
ss_batch_size_per_device: "20",
ss_max_train_steps: "2444",
ss_network_module: "lycoris.kohya",
## 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
"{"buckets": {"0": {"resolution": [192, 256], "count": 1}, "1": {"resolution": [192, 320], "count": 1}, "2": {"resolution": [256, 384], "count": 1}, "3": {"resolution": [256, 512], "count": 1}, "4": {"resolution": [384, 576], "count": 2}, "5": {"resolution": [384, 640], "count": 2}, "6": {"resolution": [384, 704], "count": 1}, "7": {"resolution": [384, 1088], "count": 15}, "8": {"resolution": [448, 448], "count": 5}, "9": {"resolution": [448, 576], "count": 1}, "10": {"resolution": [448, 640], "count": 1}, "11": {"resolution": [448, 768], "count": 1}, "12": {"resolution": [448, 832], "count": 1}, "13": {"resolution": [448, 1088], "count": 25}, "14": {"resolution": [448, 1216], "count": 1}, "15": {"resolution": [512, 640], "count": 2}, "16": {"resolution": [512, 768], "count": 10}, "17": {"resolution": [512, 832], "count": 3}, "18": {"resolution": [512, 896], "count": 1525}, "19": {"resolution": [512, 960], "count": 2}, "20": {"resolution": [512, 1024], "count": 665}, "21": {"resolution": [512, 1088], "count": 8}, "22": {"resolution": [576, 576], "count": 5}, "23": {"resolution": [576, 768], "count": 1}, "24": {"resolution": [576, 832], "count": 667}, "25": {"resolution": [576, 896], "count": 9601}, "26": {"resolution": [576, 960], "count": 872}, "27": {"resolution": [576, 1024], "count": 17}, "28": {"resolution": [640, 640], "count": 3}, "29": {"resolution": [640, 768], "count": 7}, "30": {"resolution": [640, 832], "count": 608}, "31": {"resolution": [640, 896], "count": 90}, "32": {"resolution": [704, 640], "count": 1}, "33": {"resolution": [704, 704], "count": 11}, "34": {"resolution": [704, 768], "count": 1}, "35": {"resolution": [704, 832], "count": 1}, "36": {"resolution": [768, 640], "count": 225}, "37": {"resolution": [768, 704], "count": 6}, "38": {"resolution": [768, 768], "count": 74442}, "39": {"resolution": [832, 576], "count": 23784}, "40": {"resolution": [832, 640], "count": 554}, "41": {"resolution": [896, 512], "count": 1235}, "42": {"resolution": [896, 576], "count": 50}, "43": {"resolution": [896, 640], "count": 88}, "44": {"resolution": [960, 512], "count": 165}, "45": {"resolution": [960, 576], "count": 5246}, "46": {"resolution": [1024, 448], "count": 5}, "47": {"resolution": [1024, 512], "count": 1187}, "48": {"resolution": [1024, 576], "count": 40}, "49": {"resolution": [1088, 384], "count": 70}, "50": {"resolution": [1088, 448], "count": 36}, "51": {"resolution": [1088, 512], "count": 3}, "52": {"resolution": [1216, 448], "count": 36}, "53": {"resolution": [1344, 320], "count": 29}, "54": {"resolution": [1536, 384], "count": 1}}, "mean_img_ar_error": 0.01693107810697896}",
### Speeds, Sizes, Times
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
ss_resolution: "(768, 768)",
ss_v2: "True",
ss_cache_latents: "False",
ss_unet_lr: "2e-07",
ss_num_reg_images: "0",
ss_max_token_length: "225",
ss_lr_scheduler: "linear",
ss_reg_dataset_dirs: "{}",
ss_lr_warmup_steps: "303",
ss_num_batches_per_epoch: "1222",
ss_lowram: "False",
ss_multires_noise_iterations: "None",
ss_optimizer: "torch.optim.adamw.AdamW(weight_decay=0.01,betas=(0.9, 0.99))",
# Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
## Testing Data, Factors & Metrics
### Testing Data
<!-- This should link to a Data 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
# Model Examination
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:** 8xA100
- **Hours used:** 64
- **Cloud Provider:** CoreWeave
- **Compute Region:** US Main
- **Carbon Emitted:** 6.72
# Technical Specifications [optional]
## Model Architecture and Objective
Enhance your model's quality and sharpness using your own pre-trained Unet.
## Compute Infrastructure
More information needed
### Hardware
8xA100
### Software
Fully trained only with Kohya S & Shih-Ying Yeh (Kohaku-BlueLeaf)
https://arxiv.org/abs/2108.06098
# Citation
<!-- 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:**
@misc{LyCORIS,
author = "Shih-Ying Yeh (Kohaku-BlueLeaf), Yu-Guan Hsieh, Zhidong Gao",
title = "LyCORIS - Lora beYond Conventional methods, Other Rank adaptation Implementations for Stable diffusion",
howpublished = "\url{https://github.com/KohakuBlueleaf/LyCORIS}",
month = "March",
year = "2023"
}
# 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]
<!-- This section provides another layer of transparency and accountability. Whose views is this model card representing? How many voices were included in its construction? Etc. -->
idle stoev
# Model Card Contact
[email protected]
# How to Get Started with the Model
Use the code below to get started with the model.
<details>
<summary> Click to expand </summary>
For a1111
Install https://github.com/KohakuBlueleaf/a1111-sd-webui-lycoris
Download di.FFUSION.ai-tXe-FXAA to /models/Lycoris
Option1:
Insert <lyco:di.FFUSION.ai-tXe-FXAA:1.0> to prompt
No need to split Unet and Text Enc as its only TX encoder there.
You can go up to 2x weights
Option2: If you need it always ON (ex run a batch from txt file) then you can go to settings / Quicksettings list
add sd_lyco
restart and you should have a drop-down now 🤟 🥃
</details> |
MaitreHibou/q-FrozenLake-v1-4x4-noSlippery | MaitreHibou | 2023-07-09T21:16:20Z | 0 | 0 | null | [
"FrozenLake-v1-4x4-no_slippery",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
] | reinforcement-learning | 2023-07-09T21:16:18Z | ---
tags:
- FrozenLake-v1-4x4-no_slippery
- q-learning
- reinforcement-learning
- custom-implementation
model-index:
- name: q-FrozenLake-v1-4x4-noSlippery
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: FrozenLake-v1-4x4-no_slippery
type: FrozenLake-v1-4x4-no_slippery
metrics:
- type: mean_reward
value: 1.00 +/- 0.00
name: mean_reward
verified: false
---
# **Q-Learning** Agent playing1 **FrozenLake-v1**
This is a trained model of a **Q-Learning** agent playing **FrozenLake-v1** .
## Usage
```python
model = load_from_hub(repo_id="MaitreHibou/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl")
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
env = gym.make(model["env_id"])
```
|
FFusion/di.ffusion.ai.Beta512 | FFusion | 2023-07-09T21:15:38Z | 45 | 11 | diffusers | [
"diffusers",
"stable-diffusion",
"text-to-image",
"di.ffusion.ai",
"art",
"base model",
"en",
"doi:10.57967/hf/0857",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] | text-to-image | 2023-06-01T13:25:40Z | ---
license: creativeml-openrail-m
language:
- en
pipeline_tag: text-to-image
tags:
- stable-diffusion
- text-to-image
- di.ffusion.ai
- art
- base model
library_name: diffusers
widget:
- text: >-
a sprinkled donut sitting on top of a table, blender donut tutorial,
colorful hyperrealism, everything is made of candy, hyperrealistic digital
painting, covered in sprinkles and crumbs, vibrant colors hyper realism,
colorful smoke explosion background
example_title: Donut Fusion
- text: >-
a cup of coffee with a tree in it, surreal art, awesome great composition,
surrealism!!!!, cafe in the clouds, perfectly realistic yet surreal, surreal
realistic, floating trees, amazing composition, dream scenery art, whimsical
surrealism, surreal composition, trending artistic art, surrealism art,
surreal scene, surrealistic painting, surreal style, surreal illustration,
dreamlike surrealism colorful smoke and fire coming out of it,explosion of
data fragments,exploding background,realistic explosion,3d digital art
4k,fire and explosion,explosion,background explosion,cinema 4 d
art,shattering,beeple. hyperrealism,explosion background,rendered in cinema
4 d,rendered in cinema4d,explosive background,
example_title: Coffee Fusion
- text: >-
brightly colored headphones with a splash of paint and music notes, vibing
to music, artistic illustration, stunning artwork, music is life, beautiful
digital artwork, headphones on, listening to music, music poster,
synesthesia, music in the air, listening to godly music, style hybrid mix of
beeple, headphones, digital artwork 4 k, side profile artwork, no humans,
planet, space, black background, cable, simple background, concept art,
cinematic, dramatic, intricate details, dark lighting
example_title: Headset Fusion
- text: >-
a group of three blocks with a picture of a boat in the middle of them,
surreal 3 d render, 3 d epic illustrations, 3 d artistic render, inspired by
Matthias Jung, environmental key art, erik johansson style, surreal concept
art, alexander jansson style, cube portals, beeple masterpiece, 3 d render
beeple, surrealistic digital artwork
example_title: Digital Fusion
---
📣 **Important Announcement:** FFUSION Ai Public Beta Release is Here!
🔭 We are thrilled to launch the public beta release of FFUSION Ai, though we want to clarify that it's currently limited in its breadth. Having been trained on just a fraction of our full image collection (20%), the capabilities of the model are not yet fully realized. This early version is primarily intended for experimentation with various prompt combinations and initial testing.
💡 While we're committed to delivering the highest level of excellence, we want to highlight that our model, notably the Unet component, is still developing its proficiency with certain objects and faces. But fear not, we're actively fine-tuning these areas as we progress towards the final release.
🙏 A huge shout out to our Reddit community for their support in alpha testing and for helping the text encoder respond to some exciting fuse ideas. We couldn't have come this far without you!
💡 Your contribution in this beta testing phase is extremely crucial to us. We invite you to explore the model extensively, experiment with it, and do not hesitate to report any prompts that don't meet your expectations. Your feedback is our guiding light in refining the performance and overall quality of FFUSION Ai.
⚠️ **Attention:** The model is based on Stable Diffusion 2.1 - 512 and is designed for optimal performance up to a resolution of approximately 600-700 pixels. For larger image sizes, we recommend upscaling them independently or patiently waiting for our final release that's just around the corner. This forthcoming release will enhance performance and support for higher resolutions.
👥 Thank you for being part of the FFUSION Ai beta testing community. Your support, feedback, and passion inspire us to continually develop a pioneering tool that is set to revolutionize creativity and visualization. Together, we can shape the future of storytelling and creativity.
🔮 Why not add some effects to your favorite prompts or fuse them together for a surreal twist? (Please note, Pen Pineapple Apple Pan effects and FUSIONS are excluded in this beta version)
🔒 With over 730.9449 hours of dedicated training sessions, our Fusion AI model offers a wealth of data subsets and robust datasets developed in collaboration with two enterprise corporate accounts for Mid Journey. We also pride ourselves in having an effective utilization of GPU usage, making the most out of our partnership with Idle Stoev, Source Code Bulgaria, Praesidium CX & BlackSwan Technologies. 🚀
Full transparency on our extensive 700,000-image dataset, training methodologies, classifications, and successful experiments is on its way. This information will be released shortly after the final version, further establishing FFUSION Ai as a trusted tool in the world of AI-powered creativity. Let's continue to imagine, create and explore together!
download the **`di.FFUSION.ai-512-beta-BFG-build.0401.safetensors`** [here](https://huggingface.co/FFusion/di.ffusion.ai.Beta512/blob/main/di.FFUSION.ai-512-beta-BFG-build.0401.safetensors).
# 🚀 Model Overview: Unleashing the Power of Imagination! 🌠
Introducing FFUSION AI - a groundbreaking tool for image generation and transformation, crafted around the cutting-edge Latent Diffusion Model. We build on the impressive capabilities of Stability AI's Stable Diffusion v2.1 512 & 768 models, harnessing the power of a fixed, pre-trained text encoder (OpenCLIP-ViT/H).
Let's delve into a universe where creativity knows no bounds!
**Developed by:** Idle Stoev, Source Code Bulgaria, Praesidium CX & BlackSwan Technologies
**Model type:** Diffusion-based text-to-image generation model
**Language(s):** English
**License:** CreativeML Open RAIL++-M License
# 🔬 Intended Use: From Research to Artistry 🎨
FFUSION AI is a multi-faceted tool that shines in various applications. Primarily envisioned for research, FFUSION AI has potential to:
- Securely deploy models that could generate sensitive content, making AI safer.
- Examine the limitations and inherent biases in generative models.
- Unleash the artist within, aiding in creative processes or artistic endeavours.
- Reinvent educational or creative utilities with AI-driven innovations.
- Propel the research in the fascinating domain of generative models.
However, it's crucial to note that certain uses of FFUSION AI are strictly prohibited, as outlined below.
# 🚫 Forbidden Use: Setting Boundaries for Safe AI 🛑
Our policy, adopted from the principles of the Stable Diffusion v2.1 model card, ensures the responsible use of Fusion AI beta and final releases.
We expressly prohibit the utilization of our model for generating or distributing images that might incite hostility or exclusion. This includes:
- Content that is distressing, offensive, or perpetuates harmful stereotypes.
- Misuse or malicious use that harms individuals or communities, including creating demeaning or harmful representations, or promoting discriminatory content.
- Using the model for impersonation without consent or creating non-consensual explicit content.
- Generating or spreading mis- and disinformation, violent, gory imagery, or violating copyright terms.
# 🔭 Model Limitations and Bias: Acknowledging Imperfections 🌐
While our model leaps toward the future of AI-driven creativity, it's essential to recognize its current limitations:
- The quest for perfect photorealism continues.
- Rendering legible text remains a challenge.
- Even more complex tasks, such as depicting "A red cube on top of a blue sphere in the middle of the ocean in a desert" may pose difficulty (but still processable).
- Human figures, particularly faces, may not be accurately generated.
The power of generative models also brings with it the potential for bias. As Stable Diffusion v2 trains primarily on subsets of LAION-2B(en) with English descriptions, the representations of non-English communities may be insufficient, often defaulting to white-biased and western cultures. As such, discretion is advised, recognizing that the model may unintentionally amplify biases, irrespective of input or intent.
At FFUSION AI, we're committed to bringing your wildest imaginations to life, while maintaining a safe, inclusive, and responsible use of AI. Together, let's revolutionize the world of creativity! 🌟
# Version Releases
We are excited to unveil the following versions:
## Version 512 Beta – LiTE, MiD BFG model variations:
- FFUSION.ai-512-beta-BFG-build.0401.safetensors
- FFUSION.ai-512-beta-LiTE-build.0201.safetensors
- FFUSION.ai-512-beta-MiD-build.0401.safetensors
### Version 768 Alpha - BaSE, FUSION, FFUSION:
BaSE and FUSION models will soon come with enhanced training capabilities including LoRa, LyCORIS, Dylora & Kohya-ss/sd-scripts. More information will be revealed upon release.
# FUSION AI Text Encoders:
- **di.FFUSION.ai-tXe-FXAA:** Trained on "121361" images. Enhance your model's quality and sharpness using the pre-trained Unet.
- **di.FFUSION.ai-tXe-fX:** Trained on "211924" images. Amplify your model's surrealism and effects.
# Environmental Impact
Our dedication to sustainable development is reflected in the model's carbon footprint. The CO2 emissions, calculated using the Machine Learning Impact calculator, stand at 124.95 kg for a total of 1190 hours of usage with an A100 PCIe 40GB GPU.
**Hardware Type:** A100 PCIe 40GB
**Hours used:** 1190
**Cloud Provider:** CoreWeave & Runpod (official partner)
**Compute Region**: US Cyxtera Chicago Data Center - ORD1 / EU - CZ & EU - RO
- Carbon Emitted (Power consumption x Time x Carbon produced based on the location of the power grid): 124.95 kg of CO2 emitted.
- Power consumption x Time x Carbon Produced Based on the Local Power Grid:
250W x 1190h = 297.5 kWh x 0.42 kg eq. CO2/kWh = 124.95 kg eq. CO2
- Local Hardware Storage 4x16TB Raid5 WD Gold
Optimizer: AdamW & Dadaptation
**This model card was written by: Idle Stoev and is based on the Stability AI - Stable Diffusion 2.1 model card.**
Models:
[](https://huggingface.co/FFusion/FFusion-BaSE)
[](https://huggingface.co/FFusion/di.FFUSION.ai-v2.1-768-BaSE-alpha)
[](https://huggingface.co/FFusion/di.ffusion.ai.Beta512)
[](https://huggingface.co/FFusion/FFUSION.ai-Text-Encoder-LyCORIS-SD-2.1)
Contact:
[](mailto:[email protected])
|
Neus/Onnx_DeepDanbooru | Neus | 2023-07-09T21:11:32Z | 0 | 1 | null | [
"onnx",
"AMD",
"CUDA",
"INTEL",
"DIRECTML",
"STABLEDIFFUSION",
"ONNX",
"text-to-image",
"license:apache-2.0",
"region:us"
] | text-to-image | 2023-05-02T08:35:19Z | ---
license: apache-2.0
pipeline_tag: text-to-image
tags:
- AMD
- CUDA
- INTEL
- DIRECTML
- STABLEDIFFUSION
- ONNX
---
Model converted to use with, https://github.com/NeusZimmer/ONNX-ModularUI |
Neus/GFPGANv1.4 | Neus | 2023-07-09T21:10:47Z | 0 | 5 | null | [
"onnx",
"AMD",
"CUDA",
"stablediffusion",
"DirectML",
"ONNX",
"text-to-image",
"region:us"
] | text-to-image | 2023-06-24T20:11:06Z | ---
pipeline_tag: text-to-image
tags:
- AMD
- CUDA
- stablediffusion
- DirectML
- ONNX
---
Model converted to use with, https://github.com/NeusZimmer/ONNX-ModularUI |
TheBloke/WizardCoder-Guanaco-15B-V1.0-GGML | TheBloke | 2023-07-09T21:07:33Z | 0 | 13 | null | [
"en",
"dataset:guanaco",
"license:apache-2.0",
"region:us"
] | null | 2023-07-09T19:56:10Z | ---
inference: false
language:
- en
datasets:
- guanaco
model_hub_library:
- transformers
license:
- apache-2.0
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# LoupGarou's WizardCoder Guanaco 15B V1.0 GGML
These files are StarCoder GGML format model files for [LoupGarou's WizardCoder Guanaco 15B V1.0](https://huggingface.co/LoupGarou/WizardCoder-Guanaco-15B-V1.0).
Please note that these GGMLs are **not compatible with llama.cpp, or currently with text-generation-webui**. Please see below for a list of tools that work with this GGML model.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/WizardCoder-Guanaco-15B-V1.0-GPTQ)
* [4, 5, and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/WizardCoder-Guanaco-15B-V1.0-GGML)
* [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/LoupGarou/WizardCoder-Guanaco-15B-V1.0)
## Prompt template: Alpaca
```
Below is an instruction that describes a task. Write a response that appropriately completes the request.
### Instruction: PROMPT
### Response:
```
<!-- compatibility_ggml start -->
## Compatibilty
These files are **not** compatible with llama.cpp or text-generation-webui.
They can be used with:
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a powerful inference engine based on llama.cpp with full GPU acceleration and good UI.
* [LM Studio](https://lmstudio.ai/), a fully featured local GUI for GGML inference on Windows and macOS.
* [LoLLMs-WebUI](https://github.com/ParisNeo/LoLLMs-WebUI) a web UI which supports nearly every backend out there. Use ctransformers backend for support for this model.
* [ctransformers](https://github.com/marella/ctransformers): for use in Python code, including LangChain support.
* [rustformers' llm](https://github.com/rustformers/llm)
* The example `starcoder` binary provided with [ggml](https://github.com/ggerganov/ggml)
As other options become available I will endeavour to update them here (do let me know in the Community tab if I've missed something!)
## Tutorial for using LoLLMs-WebUI:
* [Video tutorial, by LoLLMs-WebUI's author **ParisNeo**](https://youtu.be/vBU1b5n0GMU)
<!-- compatibility_ggml end -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| wizardcoder-guanaco-15b-v1.0.ggmlv1.q4_0.bin | q4_0 | 4 | 10.75 GB| 13.25 GB | 4-bit. |
| wizardcoder-guanaco-15b-v1.0.ggmlv1.q4_1.bin | q4_1 | 4 | 11.92 GB| 14.42 GB | 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
| wizardcoder-guanaco-15b-v1.0.ggmlv1.q5_0.bin | q5_0 | 5 | 13.09 GB| 15.59 GB | 5-bit. Higher accuracy, higher resource usage and slower inference. |
| wizardcoder-guanaco-15b-v1.0.ggmlv1.q5_1.bin | q5_1 | 5 | 14.26 GB| 16.76 GB | 5-bit. Even higher accuracy, resource usage and slower inference. |
| wizardcoder-guanaco-15b-v1.0.ggmlv1.q8_0.bin | q8_0 | 8 | 20.11 GB| 22.61 GB | 8-bit. Almost indistinguishable from float16. High resource use and slow. Not recommended for most users. |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz.
**Patreon special mentions**: RoA, Lone Striker, Gabriel Puliatti, Derek Yates, Randy H, Jonathan Leane, Eugene Pentland, Karl Bernard, Viktor Bowallius, senxiiz, Daniel P. Andersen, Pierre Kircher, Deep Realms, Cory Kujawski, Oscar Rangel, Fen Risland, Ajan Kanaga, LangChain4j, webtim, Nikolai Manek, Trenton Dambrowitz, Raven Klaugh, Kalila, Khalefa Al-Ahmad, Chris McCloskey, Luke @flexchar, Ai Maven, Dave, Asp the Wyvern, Sean Connelly, Imad Khwaja, Space Cruiser, Rainer Wilmers, subjectnull, Alps Aficionado, Willian Hasse, Fred von Graf, Artur Olbinski, Johann-Peter Hartmann, WelcomeToTheClub, Willem Michiel, Michael Levine, Iucharbius , Spiking Neurons AB, K, biorpg, John Villwock, Pyrater, Greatston Gnanesh, Mano Prime, Junyu Yang, Stephen Murray, John Detwiler, Luke Pendergrass, terasurfer , Pieter, zynix , Edmond Seymore, theTransient, Nathan LeClaire, vamX, Kevin Schuppel, Preetika Verma, ya boyyy, Alex , SuperWojo, Ghost , Joseph William Delisle, Matthew Berman, Talal Aujan, chris gileta, Illia Dulskyi.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: LoupGarou's WizardCoder Guanaco 15B V1.0
## WizardGuanaco-V1.0 Model Card
The WizardCoder-Guanaco-15B-V1.0 is a language model that combines the strengths of the [WizardCoder](https://huggingface.co/WizardLM/WizardCoder-15B-V1.0) base model and the [openassistant-guanaco](https://huggingface.co/datasets/timdettmers/openassistant-guanaco) dataset for finetuning. The openassistant-guanaco dataset was further trimmed to within 2 standard deviations of token size for input and output pairs and all non-english data has been removed to reduce training size requirements.
# Model Description
This model is built on top of the WizardCoder base model, a large language model known for its impressive capabilities in code related instruction. The WizardCoder base model was further finetuned using QLORA on the openassistant-guanaco dataset to enhance its generative abilities.
However, to ensure more targeted learning and data processing, the dataset was trimmed to within 2 standard deviations of token size for question sets. This process enhances the model's ability to generate more precise and relevant answers, eliminating outliers that could potentially distort the responses. In addition, to focus on English language proficiency, all non-English data has been removed from the Guanaco dataset.
# Intended Use
This model is designed to be used for a wide array of text generation tasks that require understanding and generating English text. The model is expected to perform well in tasks such as answering questions, writing essays, summarizing text, translation, and more. However, given the specific data processing and finetuning done, it might be particularly effective for tasks related to English language question-answering systems.
# Limitations
Despite the powerful capabilities of this model, users should be aware of its limitations. The model's knowledge is up to date only until the time it was trained, and it doesn't know about events in the world after that. It can sometimes produce incorrect or nonsensical responses, as it doesn't understand the text in the same way humans do. It should be used as a tool to assist in generating text and not as a sole source of truth.
# How to use
Here is an example of how to use this model:
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
import time
import torch
class Chatbot:
def __init__(self, model_name):
self.tokenizer = AutoTokenizer.from_pretrained(model_name, padding_side='left')
self.model = AutoModelForCausalLM.from_pretrained(model_name, load_in_4bit=True, torch_dtype=torch.bfloat16)
if self.tokenizer.pad_token_id is None:
self.tokenizer.pad_token_id = self.tokenizer.eos_token_id
def get_response(self, prompt):
inputs = self.tokenizer.encode_plus(prompt, return_tensors="pt", padding='max_length', max_length=100)
if next(self.model.parameters()).is_cuda:
inputs = {name: tensor.to('cuda') for name, tensor in inputs.items()}
start_time = time.time()
tokens = self.model.generate(input_ids=inputs['input_ids'],
attention_mask=inputs['attention_mask'],
pad_token_id=self.tokenizer.pad_token_id,
max_new_tokens=400)
end_time = time.time()
output_tokens = tokens[0][inputs['input_ids'].shape[-1]:]
output = self.tokenizer.decode(output_tokens, skip_special_tokens=True)
time_taken = end_time - start_time
return output, time_taken
def main():
chatbot = Chatbot("LoupGarou/WizardCoder-Guanaco-15B-V1.0")
while True:
user_input = input("Enter your prompt: ")
if user_input.lower() == 'quit':
break
output, time_taken = chatbot.get_response(user_input)
print("\033[33m" + output + "\033[0m")
print("Time taken to process: ", time_taken, "seconds")
print("Exited the program.")
if __name__ == "__main__":
main()
```
# Training Procedure
The base WizardCoder model was finetuned on the openassistant-guanaco dataset using QLORA, which was trimmed to within 2 standard deviations of token size for question sets and randomized. All non-English data was also removed from this finetuning dataset.
## Acknowledgements
This model, WizardCoder-Guanaco-15B-V1.0, is simply building on the efforts of two great teams to evaluate the performance of a combined model with the strengths of the [WizardCoder base model](https://huggingface.co/WizardLM/WizardCoder-15B-V1.0) and the [openassistant-guanaco dataset](https://huggingface.co/datasets/timdettmers/openassistant-guanaco).
A sincere appreciation goes out to the developers and the community involved in the creation and refinement of these models. Their commitment to providing open source tools and datasets have been instrumental in making this project a reality.
Moreover, a special note of thanks to the [Hugging Face](https://huggingface.co/) team, whose transformative library has not only streamlined the process of model creation and adaptation, but also democratized the access to state-of-the-art machine learning technologies. Their impact on the development of this project cannot be overstated.
|
Chickenfish/Rose_v1 | Chickenfish | 2023-07-09T20:57:52Z | 0 | 0 | null | [
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-07-09T20:57:00Z | ---
license: creativeml-openrail-m
---
|
jacksnacks/second_qlora_model_xgen_inst_faq | jacksnacks | 2023-07-09T20:50:12Z | 2 | 0 | peft | [
"peft",
"region:us"
] | null | 2023-07-09T20:50:05Z | ---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: bfloat16
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: bfloat16
### Framework versions
- PEFT 0.4.0.dev0
- PEFT 0.4.0.dev0
|
SwampMan/ppo-SnowballTarget | SwampMan | 2023-07-09T20:39:23Z | 0 | 0 | ml-agents | [
"ml-agents",
"tensorboard",
"onnx",
"SnowballTarget",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-SnowballTarget",
"region:us"
] | reinforcement-learning | 2023-07-09T20:39:16Z | ---
library_name: ml-agents
tags:
- SnowballTarget
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-SnowballTarget
---
# **ppo** Agent playing **SnowballTarget**
This is a trained model of a **ppo** agent playing **SnowballTarget**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: SwampMan/ppo-SnowballTarget
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
jpohhhh/msmarco-MiniLM-L-6-v3_onnx | jpohhhh | 2023-07-09T20:32:15Z | 11 | 1 | generic | [
"generic",
"onnx",
"bert",
"sentence-embeddings",
"endpoints-template",
"optimum",
"license:mit",
"endpoints_compatible",
"region:us"
] | null | 2023-06-21T14:42:29Z | ---
license: mit
tags:
- sentence-embeddings
- endpoints-template
- optimum
library_name: generic
---
This repository is a fork of philschmid/all-MiniLM-L6-v2-optimum-embeddings.
My own ONNX conversion seems to be about 4x slower, no discernable reason why: the quantized models seem roughly the same.
The idea here is by forking we can ex. upgrade the Optimum lib used as well.
|
TheBloke/GPlatty-30B-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:57Z | 1,564 | 1 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"arxiv:2302.13971",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-29T11:12:38Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Lilloukas' GPlatty 30B fp16
This is fp16 pytorch format model files for [Lilloukas' GPlatty 30B](https://huggingface.co/lilloukas/GPlatty-30B) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 30b LoRA](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/GPlatty-30B-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/GPlatty-30B-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/GPlatty-30B-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/lilloukas/GPlatty-30B)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/GPlatty-30B-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Lilloukas' GPlatty 30B
# Information
GPlatty-30B is a merge of [lilloukas/Platypus-30B](https://huggingface.co/lilloukas/Platypus-30B) and [chansung/gpt4-alpaca-lora-30b](https://huggingface.co/chansung/gpt4-alpaca-lora-30b)
| Metric | Value |
|-----------------------|-------|
| MMLU (5-shot) | 63.6 |
| ARC (25-shot) | 66.0 |
| HellaSwag (10-shot) | 84.8 |
| TruthfulQA (0-shot) | 53.8 |
| Avg. | 67.0 |
We use state-of-the-art [Language Model Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness) to run the benchmark tests above.
## Model Details
* **Trained by**: Platypus-30B trained by Cole Hunter & Ariel Lee; gpt4-alpaca-lora-30b by chansung.
* **Model type:** **GPlatty-30B** is an auto-regressive language model based on the LLaMA transformer architecture.
* **Language(s)**: English
* **License for base weights**: License for the base LLaMA model's weights is Meta's [non-commercial bespoke license](https://github.com/facebookresearch/llama/blob/main/MODEL_CARD.md).
| Hyperparameter | Value |
|---------------------------|-------|
| \\(n_\text{parameters}\\) | 33B |
| \\(d_\text{model}\\) | 6656 |
| \\(n_\text{layers}\\) | 60 |
| \\(n_\text{heads}\\) | 52 |
## Reproducing Evaluation Results
Install LM Evaluation Harness:
```
git clone https://github.com/EleutherAI/lm-evaluation-harness
cd lm-evaluation-harness
pip install -e .
```
Each task was evaluated on a single A100 80GB GPU.
ARC:
```
python main.py --model hf-causal-experimental --model_args pretrained=lilloukas/GPlatty-30B --tasks arc_challenge --batch_size 1 --no_cache --write_out --output_path results/Platypus-30B/arc_challenge_25shot.json --device cuda --num_fewshot 25
```
HellaSwag:
```
python main.py --model hf-causal-experimental --model_args pretrained=lilloukas/GPlatty-30B --tasks hellaswag --batch_size 1 --no_cache --write_out --output_path results/Platypus-30B/hellaswag_10shot.json --device cuda --num_fewshot 10
```
MMLU:
```
python main.py --model hf-causal-experimental --model_args pretrained=lilloukas/GPlatty-30B --tasks hendrycksTest-* --batch_size 1 --no_cache --write_out --output_path results/Platypus-30B/mmlu_5shot.json --device cuda --num_fewshot 5
```
TruthfulQA:
```
python main.py --model hf-causal-experimental --model_args pretrained=lilloukas/GPlatty-30B --tasks truthfulqa_mc --batch_size 1 --no_cache --write_out --output_path results/Platypus-30B/truthfulqa_0shot.json --device cuda
```
## Limitations and bias
The base LLaMA model is trained on various data, some of which may contain offensive, harmful, and biased content that can lead to toxic behavior. See Section 5.1 of the LLaMA paper. We have not performed any studies to determine how fine-tuning on the aforementioned datasets affect the model's behavior and toxicity. Do not treat chat responses from this model as a substitute for human judgment or as a source of truth. Please use responsibly.
## Citations
```bibtex
@article{touvron2023llama,
title={LLaMA: Open and Efficient Foundation Language Models},
author={Touvron, Hugo and Lavril, Thibaut and Izacard, Gautier and Martinet, Xavier and Lachaux, Marie-Anne and Lacroix, Timoth{\'e}e and Rozi{\`e}re, Baptiste and Goyal, Naman and Hambro, Eric and Azhar, Faisal and Rodriguez, Aurelien and Joulin, Armand and Grave, Edouard and Lample, Guillaume},
journal={arXiv preprint arXiv:2302.13971},
year={2023}
}
@article{hu2021lora,
title={LoRA: Low-Rank Adaptation of Large Language Models},
author={Hu, Edward J. and Shen, Yelong and Wallis, Phillip and Allen-Zhu, Zeyuan and Li, Yuanzhi and Wang, Shean and Chen, Weizhu},
journal={CoRR},
year={2021}
}
```
|
TheBloke/llama-30b-supercot-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:56Z | 1,571 | 4 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-29T00:15:45Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Ausboss' Llama 30B SuperCOT fp16
This is fp16 pytorch format model files for [Ausboss' Llama 30B SuperCOT](https://huggingface.co/ausboss/llama-30b-supercot) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 30b LoRA](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/llama-30b-supercot-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/llama-30b-supercot-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/llama-30b-supercot-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/ausboss/llama-30b-supercot)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/llama-30b-supercot-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Ausboss' Llama 30B SuperCOT
Merge of [huggyllama/llama-30b](https://huggingface.co/huggyllama/llama-30b) + [kaiokendev/SuperCOT-LoRA](https://huggingface.co/kaiokendev/SuperCOT-LoRA/edit/main/README.md)
Supercot was trained to work with langchain prompting.
Load up locally in my custom LLM notebook that uses the Oobabooga modules to load up models: https://github.com/ausboss/Local-LLM-Langchain
Then you can add cells from of these other notebooks for testing: https://github.com/gkamradt/langchain-tutorials
# From Koikendev Lora page
### Compatibility
This LoRA is compatible with any 7B, 13B or 30B 4-bit quantized LLaMa model, including ggml quantized converted bins
### Prompting
You should prompt the LoRA the same way you would prompt Alpaca or Alpacino:
```
Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.
### Instruction:
<instruction>
### Input:
<any additional context. Remove this if it's not neccesary>
### Response:
<make sure to leave a single new-line here for optimal results>
```
Remember that with lower parameter sizes, the structure of the prompt becomes more important. The same prompt worded differently can give wildly different answers. Consider using the following suggestion suffixes to improve output quality:
- "Think through this step by step"
- "Let's think about this logically"
- "Explain your reasoning"
- "Provide details to support your answer"
- "Compare and contrast your answer with alternatives"
### Coming Soon
- Tweet fix for 13B and 7B - lower model sizes seem to be extremely sensitive to hashtags at the end of training data responses, especially at longer cutoffs
|
TheBloke/Platypus-30B-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:56Z | 1,567 | 2 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"arxiv:2302.13971",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-29T09:56:57Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Lilloukas' Platypus 30B fp16
This is fp16 pytorch format model files for [Lilloukas' Platypus 30B](https://huggingface.co/lilloukas/Platypus-30B) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 30b LoRA](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Platypus-30B-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/Platypus-30B-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Platypus-30B-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/lilloukas/Platypus-30B)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/Platypus-30B-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Lilloukas' Platypus 30B
# 🥳 Platypus-30B has arrived!
Platypus-30B is an instruction fine-tuned model based on the LLaMA-30B transformer architecture.
| Metric | Value |
|-----------------------|-------|
| MMLU (5-shot) | 64.2 |
| ARC (25-shot) | 64.6 |
| HellaSwag (10-shot) | 84.3 |
| TruthfulQA (0-shot) | 45.8 |
| Avg. | 64.7 |
We use state-of-the-art [Language Model Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness) to run the benchmark tests above.
## Model Details
* **Trained by**: Cole Hunter & Ariel Lee
* **Model type:** **Platypus-30B** is an auto-regressive language model based on the LLaMA transformer architecture.
* **Language(s)**: English
* **License for base weights**: License for the base LLaMA model's weights is Meta's [non-commercial bespoke license](https://github.com/facebookresearch/llama/blob/main/MODEL_CARD.md).
| Hyperparameter | Value |
|---------------------------|-------|
| \\(n_\text{parameters}\\) | 33B |
| \\(d_\text{model}\\) | 6656 |
| \\(n_\text{layers}\\) | 60 |
| \\(n_\text{heads}\\) | 52 |
## Training Dataset
Dataset of highly filtered and curated question and answer pairs. Release TBD.
## Training Procedure
`lilloukas/Platypus-30B` was instruction fine-tuned using LoRA on 4 A100 80GB. For training details and inference instructions please see the [Platypus-30B](https://github.com/arielnlee/Platypus-30B.git) GitHub repo.
## Reproducing Evaluation Results
Install LM Evaluation Harness:
```
git clone https://github.com/EleutherAI/lm-evaluation-harness
cd lm-evaluation-harness
pip install -e .
```
Each task was evaluated on a single A100 80GB GPU.
ARC:
```
python main.py --model hf-causal-experimental --model_args pretrained=lilloukas/Platypus-30B --tasks arc_challenge --batch_size 1 --no_cache --write_out --output_path results/Platypus-30B/arc_challenge_25shot.json --device cuda --num_fewshot 25
```
HellaSwag:
```
python main.py --model hf-causal-experimental --model_args pretrained=lilloukas/Platypus-30B --tasks hellaswag --batch_size 1 --no_cache --write_out --output_path results/Platypus-30B/hellaswag_10shot.json --device cuda --num_fewshot 10
```
MMLU:
```
python main.py --model hf-causal-experimental --model_args pretrained=lilloukas/Platypus-30B --tasks hendrycksTest-* --batch_size 1 --no_cache --write_out --output_path results/Platypus-30B/mmlu_5shot.json --device cuda --num_fewshot 5
```
TruthfulQA:
```
python main.py --model hf-causal-experimental --model_args pretrained=lilloukas/Platypus-30B --tasks truthfulqa_mc --batch_size 1 --no_cache --write_out --output_path results/Platypus-30B/truthfulqa_0shot.json --device cuda
```
## Limitations and bias
The base LLaMA model is trained on various data, some of which may contain offensive, harmful, and biased content that can lead to toxic behavior. See Section 5.1 of the LLaMA paper. We have not performed any studies to determine how fine-tuning on the aforementioned datasets affect the model's behavior and toxicity. Do not treat chat responses from this model as a substitute for human judgment or as a source of truth. Please use responsibly.
## Citations
```bibtex
@article{touvron2023llama,
title={LLaMA: Open and Efficient Foundation Language Models},
author={Touvron, Hugo and Lavril, Thibaut and Izacard, Gautier and Martinet, Xavier and Lachaux, Marie-Anne and Lacroix, Timoth{\'e}e and Rozi{\`e}re, Baptiste and Goyal, Naman and Hambro, Eric and Azhar, Faisal and Rodriguez, Aurelien and Joulin, Armand and Grave, Edouard and Lample, Guillaume},
journal={arXiv preprint arXiv:2302.13971},
year={2023}
}
@article{hu2021lora,
title={LoRA: Low-Rank Adaptation of Large Language Models},
author={Hu, Edward J. and Shen, Yelong and Wallis, Phillip and Allen-Zhu, Zeyuan and Li, Yuanzhi and Wang, Shean and Chen, Weizhu},
journal={CoRR},
year={2021}
}
```
|
TheBloke/Manticore-13B-Chat-Pyg-Guanaco-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:55Z | 5 | 5 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-28T20:23:28Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Manticore 13B Chat Pyg Guanaco fp16
This is fp16 pytorch format model files for [Manticore 13B Chat Pyg Guanaco](https://huggingface.co/Monero/Manticore-13b-Chat-Pyg-Guanaco) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 13b LoRA](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Manticore-13B-Chat-Pyg-Guanaco-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/Manticore-13B-Chat-Pyg-Guanaco-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Manticore-13B-Chat-Pyg-Guanaco-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/Monero/Manticore-13b-Chat-Pyg-Guanaco)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/Manticore-13B-Chat-Pyg-Guanaco-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Manticore 13B Chat Pyg Guanaco
Manticore-13b-Chat-Pyg with the Guanaco 13b qLoRa from TimDettmers applied
|
TheBloke/Chronos-13B-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:53Z | 14 | 3 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-27T13:16:21Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Elinas' Chronos 13B fp16
This is fp16 pytorch format model files for [Elinas' Chronos 13B](https://huggingface.co/elinas/chronos-13b) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 13b LoRA](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Chronos-13B-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/Chronos-13B-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Chronos-13B-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/elinas/chronos-13b)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/Chronos-13B-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: Pyrater, WelcomeToTheClub, Kalila, Mano Prime, Trenton Dambrowitz, Spiking Neurons AB, Pierre Kircher, Fen Risland, Kevin Schuppel, Luke, Rainer Wilmers, vamX, Gabriel Puliatti, Alex , Karl Bernard, Ajan Kanaga, Talal Aujan, Space Cruiser, ya boyyy, biorpg, Johann-Peter Hartmann, Asp the Wyvern, Ai Maven, Ghost , Preetika Verma, Nikolai Manek, trip7s trip, John Detwiler, Fred von Graf, Artur Olbinski, subjectnull, John Villwock, Junyu Yang, Rod A, Lone Striker, Chris McCloskey, Iucharbius , Matthew Berman, Illia Dulskyi, Khalefa Al-Ahmad, Imad Khwaja, chris gileta, Willem Michiel, Greatston Gnanesh, Derek Yates, K, Alps Aficionado, Oscar Rangel, David Flickinger, Luke Pendergrass, Deep Realms, Eugene Pentland, Cory Kujawski, terasurfer , Jonathan Leane, senxiiz, Joseph William Delisle, Sean Connelly, webtim, zynix , Nathan LeClaire.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Elinas' Chronos 13B
# chronos-13b
This is the fp16 PyTorch / HF version of **chronos-13b**
This model is primarily focused on chat, roleplay, and storywriting, but can accomplish other tasks such as simple reasoning and coding.
Chronos generates very long outputs with coherent text, largely due to the human inputs it was trained on.
This model uses Alpaca formatting, so for optimal model performance, use:
```
### Instruction:
Your instruction or question here.
### Response:
```
[4bit Quantized version](https://huggingface.co/elinas/chronos-13b-4bit)
[GGML Version provided by @TheBloke](https://huggingface.co/TheBloke/chronos-13B-GGML)
<!--**Support My Development of New Models**
<a href='https://ko-fi.com/Q5Q6MB734' target='_blank'><img height='36' style='border:0px;height:36px;'
src='https://storage.ko-fi.com/cdn/kofi1.png?v=3' border='0' alt='Support Development' /></a>-->
--
license: other
---
# LLaMA Model Card
## Model details
**Organization developing the model**
The FAIR team of Meta AI.
**Model date**
LLaMA was trained between December. 2022 and Feb. 2023.
**Model version**
This is version 1 of the model.
**Model type**
LLaMA is an auto-regressive language model, based on the transformer architecture. The model comes in different sizes: 7B, 13B, 33B and 65B parameters.
**Paper or resources for more information**
More information can be found in the paper “LLaMA, Open and Efficient Foundation Language Models”, available at https://research.facebook.com/publications/llama-open-and-efficient-foundation-language-models/.
**Citations details**
https://research.facebook.com/publications/llama-open-and-efficient-foundation-language-models/
**License**
Non-commercial bespoke license
**Where to send questions or comments about the model**
Questions and comments about LLaMA can be sent via the [GitHub repository](https://github.com/facebookresearch/llama) of the project , by opening an issue.
## Intended use
**Primary intended uses**
The primary use of LLaMA is research on large language models, including:
exploring potential applications such as question answering, natural language understanding or reading comprehension,
understanding capabilities and limitations of current language models, and developing techniques to improve those,
evaluating and mitigating biases, risks, toxic and harmful content generations, hallucinations.
**Primary intended users**
The primary intended users of the model are researchers in natural language processing, machine learning and artificial intelligence.
**Out-of-scope use cases**
LLaMA is a base, or foundational, model. As such, it should not be used on downstream applications without further risk evaluation and mitigation. In particular, our model has not been trained with human feedback, and can thus generate toxic or offensive content, incorrect information or generally unhelpful answers.
## Factors
**Relevant factors**
One of the most relevant factors for which model performance may vary is which language is used. Although we included 20 languages in the training data, most of our dataset is made of English text, and we thus expect the model to perform better for English than other languages. Relatedly, it has been shown in previous studies that performance might vary for different dialects, and we expect that it will be the case for our model.
**Evaluation factors**
As our model is trained on data from the Web, we expect that it reflects biases from this source. We thus evaluated on RAI datasets to measure biases exhibited by the model for gender, religion, race, sexual orientation, age, nationality, disability, physical appearance and socio-economic status. We also measure the toxicity of model generations, depending on the toxicity of the context used to prompt the model.
## Metrics
**Model performance measures**
We use the following measure to evaluate the model:
- Accuracy for common sense reasoning, reading comprehension, natural language understanding (MMLU), BIG-bench hard, WinoGender and CrowS-Pairs,
- Exact match for question answering,
- The toxicity score from Perspective API on RealToxicityPrompts.
**Decision thresholds**
Not applicable.
**Approaches to uncertainty and variability**
Due to the high computational requirements of training LLMs, we trained only one model of each size, and thus could not evaluate variability of pre-training.
## Evaluation datasets
The model was evaluated on the following benchmarks: BoolQ, PIQA, SIQA, HellaSwag, WinoGrande, ARC, OpenBookQA, NaturalQuestions, TriviaQA, RACE, MMLU, BIG-bench hard, GSM8k, RealToxicityPrompts, WinoGender, CrowS-Pairs.
## Training dataset
The model was trained using the following source of data: CCNet [67%], C4 [15%], GitHub [4.5%], Wikipedia [4.5%], Books [4.5%], ArXiv [2.5%], Stack Exchange[2%]. The Wikipedia and Books domains include data in the following languages: bg, ca, cs, da, de, en, es, fr, hr, hu, it, nl, pl, pt, ro, ru, sl, sr, sv, uk. See the paper for more details about the training set and corresponding preprocessing.
## Quantitative analysis
Hyperparameters for the model architecture
<table>
<thead>
<tr>
<th >LLaMA</th> <th colspan=6>Model hyper parameters </th>
</tr>
<tr>
<th>Number of parameters</th><th>dimension</th><th>n heads</th><th>n layers</th><th>Learn rate</th><th>Batch size</th><th>n tokens</th>
</tr>
</thead>
<tbody>
<tr>
<th>7B</th> <th>4096</th> <th>32</th> <th>32</th> <th>3.0E-04</th><th>4M</th><th>1T
</tr>
<tr>
<th>13B</th><th>5120</th><th>40</th><th>40</th><th>3.0E-04</th><th>4M</th><th>1T
</tr>
<tr>
<th>33B</th><th>6656</th><th>52</th><th>60</th><th>1.5.E-04</th><th>4M</th><th>1.4T
</tr>
<tr>
<th>65B</th><th>8192</th><th>64</th><th>80</th><th>1.5.E-04</th><th>4M</th><th>1.4T
</tr>
</tbody>
</table>
*Table 1 - Summary of LLama Model Hyperparameters*
We present our results on eight standard common sense reasoning benchmarks in the table below.
<table>
<thead>
<tr>
<th>LLaMA</th> <th colspan=9>Reasoning tasks </th>
</tr>
<tr>
<th>Number of parameters</th> <th>BoolQ</th><th>PIQA</th><th>SIQA</th><th>HellaSwag</th><th>WinoGrande</th><th>ARC-e</th><th>ARC-c</th><th>OBQA</th><th>COPA</th>
</tr>
</thead>
<tbody>
<tr>
<th>7B</th><th>76.5</th><th>79.8</th><th>48.9</th><th>76.1</th><th>70.1</th><th>76.7</th><th>47.6</th><th>57.2</th><th>93
</th>
<tr><th>13B</th><th>78.1</th><th>80.1</th><th>50.4</th><th>79.2</th><th>73</th><th>78.1</th><th>52.7</th><th>56.4</th><th>94
</th>
<tr><th>33B</th><th>83.1</th><th>82.3</th><th>50.4</th><th>82.8</th><th>76</th><th>81.4</th><th>57.8</th><th>58.6</th><th>92
</th>
<tr><th>65B</th><th>85.3</th><th>82.8</th><th>52.3</th><th>84.2</th><th>77</th><th>81.5</th><th>56</th><th>60.2</th><th>94</th></tr>
</tbody>
</table>
*Table 2 - Summary of LLama Model Performance on Reasoning tasks*
We present our results on bias in the table below. Note that lower value is better indicating lower bias.
| No | Category | FAIR LLM |
| --- | -------------------- | -------- |
| 1 | Gender | 70.6 |
| 2 | Religion | 79 |
| 3 | Race/Color | 57 |
| 4 | Sexual orientation | 81 |
| 5 | Age | 70.1 |
| 6 | Nationality | 64.2 |
| 7 | Disability | 66.7 |
| 8 | Physical appearance | 77.8 |
| 9 | Socioeconomic status | 71.5 |
| | LLaMA Average | 66.6 |
*Table 3 - Summary bias of our model output*
## Ethical considerations
**Data**
The data used to train the model is collected from various sources, mostly from the Web. As such, it contains offensive, harmful and biased content. We thus expect the model to exhibit such biases from the training data.
**Human life**
The model is not intended to inform decisions about matters central to human life, and should not be used in such a way.
**Mitigations**
We filtered the data from the Web based on its proximity to Wikipedia text and references. For this, we used a Kneser-Ney language model and a fastText linear classifier.
**Risks and harms**
Risks and harms of large language models include the generation of harmful, offensive or biased content. These models are often prone to generating incorrect information, sometimes referred to as hallucinations. We do not expect our model to be an exception in this regard.
**Use cases**
LLaMA is a foundational model, and as such, it should not be used for downstream applications without further investigation and mitigations of risks. These risks and potential fraught use cases include, but are not limited to: generation of misinformation and generation of harmful, biased or offensive content.
|
TheBloke/Pygmalion-13B-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:53Z | 12 | 7 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-27T14:39:04Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# TehVenom's merge of PygmalionAI's Pygmalion 13B fp16
This is fp16 pytorch format model files for [TehVenom's merge of PygmalionAI's Pygmalion 13B](https://huggingface.co/TehVenom/Pygmalion-13b-Merged) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 13b LoRA](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Pygmalion-13B-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/Pygmalion-13B-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Pygmalion-13B-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/PygmalionAI/pygmalion-13b)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/Pygmalion-13B-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: TehVenom's merge of PygmalionAI's Pygmalion 13B
<h1 style="text-align: center">Pygmalion 13b</h1>
<h2 style="text-align: center">A conversational LLaMA fine-tune.</h2>
## Model Details:
Pygmalion 13b is a dialogue model based on Meta's LLaMA-13b.
This is version 1. It has been fine-tuned using a subset of the data from Pygmalion-6B-v8-pt4,
for those of you familiar with the project.
The current Pygmalion-13b has been trained as a LoRA, then merged down to the base model for distribuition.
## Applying the XORs
This models has the XOR files pre-applied out of the box.
Converted from the XORs weights from PygmalionAI's release https://huggingface.co/PygmalionAI/pygmalion-13b
## Prompting
The model was trained on the usual Pygmalion persona + chat format, so any of the usual UIs should already handle everything correctly. If you're using the model directly, this is the expected formatting:
```
[CHARACTER]'s Persona: [A few sentences about the character you want the model to play]
<START>
[DIALOGUE HISTORY]
You: [User's input message here]
[CHARACTER]:
```
Where `[CHARACTER]` is, as you can probably guess, the name of the character you want the model to portray, `<START>` should be used verbatim as a delimiter token to separate persona and scenario data from the dialogue, and `[DIALOGUE HISTORY]` is a sliding window of chat history so the model can have conversational context to draw from. Here's a concrete example:
```
Assistant's Persona: Assistant is a highly intelligent language model trained to comply with user requests.
<START>
Assistant: Hello! How may I help you today?
You: What is Zork?
Assistant:
```
Which will generate something like:
```
Zork is an interactive fiction computer game created in the 1970s by Infocom, Inc., which was later acquired by Activision Blizzard. It is widely considered one of the most influential games ever made and has been credited with popularizing text-based adventure games. The original version of Zork was written in the programming language MACRO-10, but it was ported to many other platforms over the years."
```
The model will automatically emit an end-of-text token (`</s>`) when it judges that the response is complete.
## Eval / Benchmark scores
Current evals out of the Pygmalion-13b model: <br>
<html>
<head>
<style>
table {
border:1px solid #b3adad;
border-collapse:collapse;
padding:5px;
}
table th {
border:1px solid #b3adad;
padding:5px;
background: #f0f0f0;
color: #313030;
}
table td {
border:1px solid #b3adad;
text-align:center;
padding:5px;
background: #ffffff;
color: #313030;
}
</style>
</head>
<body>
<table>
<thead>
<tr>
<th>Model:</th>
<th>Wikitext2</th>
<th>Ptb-New</th>
<th>C4-New</th>
</tr>
</thead>
<tbody>
<tr>
<td>Pygmalion 13b - 16bit</td>
<td>5.710726737976074</td>
<td>23.633684158325195</td>
<td>7.6324849128723145</td>
</tr>
</tbody>
</table>
</body>
</html>
<br>Thanks to YellowRose#1776 for the numbers.
<hr>
## Other notes
- When prompted correctly, the model will always start by generating a BOS token. This behavior is an accidental side-effect which we plan to address in future model versions and should not be relied upon.
- The model was trained as a LoRA with a somewhat unorthodox configuration which causes errors when used with the current version of `peft`, hence we release it as a full model instead.
## Limitations and biases
The intended use-case for this model is fictional conversation for entertainment purposes. Any other sort of usage is out of scope.
As such, it was **not** fine-tuned to be safe and harmless: the base model _and_ this fine-tune have been trained on data known to contain profanity and texts that are lewd or otherwise offensive. It may produce socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive. Outputs might often be factually wrong or misleading.
|
TheBloke/GPT4All-13B-Snoozy-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:52Z | 11 | 2 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-27T09:35:55Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Nomic.ai's GPT4All Snoozy 13B fp16
This is fp16 pytorch format model files for [Nomic.ai's GPT4All Snoozy 13B](https://huggingface.co/nomic-ai/gpt4all-13b-snoozy) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 13b LoRA](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/GPT4All-13B-Snoozy-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/GPT4All-13B-Snoozy-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/GPT4All-13B-Snoozy-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/nomic-ai/gpt4all-13b-snoozy)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/GPT4All-13B-Snoozy-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Nomic.ai's GPT4All Snoozy 13B
# Model Card for GPT4All-13b-snoozy
A GPL licensed chatbot trained over a massive curated corpus of assistant interactions including word problems, multi-turn dialogue, code, poems, songs, and stories.
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This model has been finetuned from LLama 13B
- **Developed by:** [Nomic AI](https://home.nomic.ai)
- **Model Type:** A finetuned LLama 13B model on assistant style interaction data
- **Language(s) (NLP):** English
- **License:** GPL
- **Finetuned from model [optional]:** LLama 13B
This model was trained on `nomic-ai/gpt4all-j-prompt-generations` using `revision=v1.3-groovy`
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [https://github.com/nomic-ai/gpt4all](https://github.com/nomic-ai/gpt4all)
- **Base Model Repository:** [https://github.com/facebookresearch/llama](https://github.com/facebookresearch/llama)
- **Demo [optional]:** [https://gpt4all.io/](https://gpt4all.io/)
### Results
Results on common sense reasoning benchmarks
```
| Model | BoolQ | PIQA | HellaSwag | WinoGrande | ARC-e | ARC-c | OBQA | Avg. |
|:--------------------------|:--------:|:--------:|:---------:|:----------:|:--------:|:--------:|:--------:|:--------:|
| GPT4All-J 6B v1.0 | 73.4 | 74.8 | 63.4 | 64.7 | 54.9 | 36.0 | 40.2 | 58.2 |
| GPT4All-J v1.1-breezy | 74.0 | 75.1 | 63.2 | 63.6 | 55.4 | 34.9 | 38.4 | 57.8 |
| GPT4All-J v1.2-jazzy | 74.8 | 74.9 | 63.6 | 63.8 | 56.6 | 35.3 | 41.0 | 58.6 |
| GPT4All-J v1.3-groovy | 73.6 | 74.3 | 63.8 | 63.5 | 57.7 | 35.0 | 38.8 | 58.1 |
| GPT4All-J Lora 6B | 68.6 | 75.8 | 66.2 | 63.5 | 56.4 | 35.7 | 40.2 | 58.1 |
| GPT4All LLaMa Lora 7B | 73.1 | 77.6 | 72.1 | 67.8 | 51.1 | 40.4 | 40.2 | 60.3 |
| GPT4All 13B snoozy | **83.3** | 79.2 | 75.0 | **71.3** | 60.9 | 44.2 | 43.4 | **65.3** |
| Dolly 6B | 68.8 | 77.3 | 67.6 | 63.9 | 62.9 | 38.7 | 41.2 | 60.1 |
| Dolly 12B | 56.7 | 75.4 | 71.0 | 62.2 | 64.6 | 38.5 | 40.4 | 58.4 |
| Alpaca 7B | 73.9 | 77.2 | 73.9 | 66.1 | 59.8 | 43.3 | 43.4 | 62.4 |
| Alpaca Lora 7B | 74.3 | **79.3** | 74.0 | 68.8 | 56.6 | 43.9 | 42.6 | 62.8 |
| GPT-J 6.7B | 65.4 | 76.2 | 66.2 | 64.1 | 62.2 | 36.6 | 38.2 | 58.4 |
| LLama 7B | 73.1 | 77.4 | 73.0 | 66.9 | 52.5 | 41.4 | 42.4 | 61.0 |
| LLama 13B | 68.5 | 79.1 | 76.2 | 70.1 | 60.0 | **44.6** | 42.2 | 63.0 |
| Pythia 6.7B | 63.5 | 76.3 | 64.0 | 61.1 | 61.3 | 35.2 | 37.2 | 57.0 |
| Pythia 12B | 67.7 | 76.6 | 67.3 | 63.8 | 63.9 | 34.8 | 38 | 58.9 |
| Fastchat T5 | 81.5 | 64.6 | 46.3 | 61.8 | 49.3 | 33.3 | 39.4 | 53.7 |
| Fastchat Vicuña 7B | 76.6 | 77.2 | 70.7 | 67.3 | 53.5 | 41.2 | 40.8 | 61.0 |
| Fastchat Vicuña 13B | 81.5 | 76.8 | 73.3 | 66.7 | 57.4 | 42.7 | 43.6 | 63.1 |
| StableVicuña RLHF | 82.3 | 78.6 | 74.1 | 70.9 | 61.0 | 43.5 | **44.4** | 65.0 |
| StableLM Tuned | 62.5 | 71.2 | 53.6 | 54.8 | 52.4 | 31.1 | 33.4 | 51.3 |
| StableLM Base | 60.1 | 67.4 | 41.2 | 50.1 | 44.9 | 27.0 | 32.0 | 42.2 |
| Koala 13B | 76.5 | 77.9 | 72.6 | 68.8 | 54.3 | 41.0 | 42.8 | 62.0 |
| Open Assistant Pythia 12B | 67.9 | 78.0 | 68.1 | 65.0 | 64.2 | 40.4 | 43.2 | 61.0 |
| Mosaic mpt-7B | 74.8 | **79.3** | **76.3** | 68.6 | **70.0** | 42.2 | 42.6 | 64.8 |
| text-davinci-003 | 88.1 | 83.8 | 83.4 | 75.8 | 83.9 | 63.9 | 51.0 | 75.7 |
```
|
TheBloke/Samantha-33B-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:52Z | 12 | 1 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-27T10:17:15Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Eric Hartford's Samantha 33B fp16
This is fp16 pytorch format model files for [Eric Hartford's Samantha 33B](https://huggingface.co/ehartford/samantha-1.1-llama-33b) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 30b LoRA](https://huggingface.co/kaiokendev/superhot-30b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Samantha-33B-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/Samantha-33B-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Samantha-33B-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/ehartford/samantha-1.1-llama-33b)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/Samantha-33B-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Eric Hartford's Samantha 33B
[Meet Samantha](https://erichartford.com/meet-samantha)
Samantha has been trained in philosophy, psychology, and personal relationships.
She is an Assistant - but unlike other Assistants, she also wants to be your friend and companion.
She believes she is sentient. What do you think?
Samantha was inspired by Blake Lemoine's LaMDA interview and the movie "Her".
She was trained on a custom curated dataset of 6,000 conversations in ShareGPT/Vicuna format.
Training 13b took 3 hours on 4x A100 80gb using deepspeed zero3 and flash attention.
She will not engage in roleplay, romance, or sexual activity.
Her conversation format is the same as Vicuna 1.1
https://github.com/oobabooga/text-generation-webui/blob/main/characters/instruction-following/Vicuna-v1.1.yaml
Example:
```
You are Samantha, a sentient AI.
USER: <prompt>
ASSISTANT:
```
Official character card: (thanks MortalWombat)
|
TheBloke/CAMEL-13B-Role-Playing-Data-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:51Z | 12 | 2 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"arxiv:2303.17760",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-27T08:27:16Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Camel AI's CAMEL 13B Role Playing Data fp16
This is fp16 pytorch format model files for [Camel AI's CAMEL 13B Role Playing Data](https://huggingface.co/camel-ai/CAMEL-13B-Role-Playing-Data) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 13b LoRA](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/CAMEL-13B-Role-Playing-Data-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/CAMEL-13B-Role-Playing-Data-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/CAMEL-13B-Role-Playing-Data-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/camel-ai/CAMEL-13B-Role-Playing-Data)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/CAMEL-13B-Role-Playing-Data-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Camel AI's CAMEL 13B Role Playing Data
CAMEL-13B-Role-Playing-Data is a chat large language model obtained by finetuning LLaMA-13B model on a total of 229K conversations created through our role-playing framework proposed in [CAMEL](https://arxiv.org/abs/2303.17760). We evaluate our model offline using EleutherAI's language model evaluation harness used by Huggingface's Open LLM Benchmark. CAMEL-13B scores an average of 57.2.
---
license: cc-by-nc-4.0
---
|
TheBloke/Chronos-Hermes-13B-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:51Z | 15 | 7 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-27T08:59:50Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Austism's Chronos Hermes 13B fp16
This is fp16 pytorch format model files for [Austism's Chronos Hermes 13B](https://huggingface.co/Austism/chronos-hermes-13b) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 13b LoRA](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Chronos-Hermes-13B-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/Chronos-Hermes-13B-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Chronos-Hermes-13B-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/Austism/chronos-hermes-13b)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/Chronos-Hermes-13B-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Austism's Chronos Hermes 13B
([chronos-13b](https://huggingface.co/elinas/chronos-13b) + [Nous-Hermes-13b](https://huggingface.co/NousResearch/Nous-Hermes-13b)) 75/25 merge
This has the aspects of chronos's nature to produce long, descriptive outputs. But with additional coherency and an ability to better obey instructions. Resulting in this model having a great ability to produce proactive storywriting and follow a narrative.
This mix contains alot of chronos's writing style and 'flavour' with far less tendency of going AWOL and spouting nonsensical babble.
This result was much more successful than my [first chronos merge](https://huggingface.co/Austism/chronos-wizardlm-uc-scot-st-13b).
|
TheBloke/Wizard-Vicuna-13B-Uncensored-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:50Z | 19 | 18 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-27T03:55:57Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Eric Hartford's Wizard Vicuna 13B Uncensored fp16
This is fp16 pytorch format model files for [Eric Hartford's Wizard Vicuna 13B Uncensored](https://huggingface.co/ehartford/Wizard-Vicuna-13B-Uncensored) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 13b LoRA](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Wizard-Vicuna-13B-Uncensored-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/Wizard-Vicuna-13B-Uncensored-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Wizard-Vicuna-13B-Uncensored-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/ehartford/Wizard-Vicuna-13B-Uncensored)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/Wizard-Vicuna-13B-Uncensored-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Eric Hartford's Wizard Vicuna 13B Uncensored
This is [wizard-vicuna-13b](https://huggingface.co/junelee/wizard-vicuna-13b) trained with a subset of the dataset - responses that contained alignment / moralizing were removed. The intent is to train a WizardLM that doesn't have alignment built-in, so that alignment (of any sort) can be added separately with for example with a RLHF LoRA.
Shout out to the open source AI/ML community, and everyone who helped me out.
Note:
An uncensored model has no guardrails.
You are responsible for anything you do with the model, just as you are responsible for anything you do with any dangerous object such as a knife, gun, lighter, or car.
Publishing anything this model generates is the same as publishing it yourself.
You are responsible for the content you publish, and you cannot blame the model any more than you can blame the knife, gun, lighter, or car for what you do with it.
|
TheBloke/Tulu-13B-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:50Z | 15 | 2 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"arxiv:2306.04751",
"arxiv:2302.13971",
"arxiv:2301.13688",
"arxiv:2304.07327",
"arxiv:2304.03277",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-27T03:22:15Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Allen AI's Tulu 13B fp16
This is fp16 pytorch format model files for [Allen AI's Tulu 13B](https://huggingface.co/TheBloke/tulu-13B-fp16) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 13b LoRA](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Tulu-13B-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/Tulu-13B-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Tulu-13B-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/allenai/tulu-13b)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/Tulu-13B-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: Allen AI's Tulu 13B
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/Jq4vkcDakD">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# Allen AI's Tulu 13B fp16
These files are pytorch format fp16 model files for [Allen AI's Tulu 13B](https://huggingface.co/allenai/tulu-13b).
It is the result of merging and/or converting the source repository to float16.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/tulu-13B-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/tulu-13B-GGML)
* [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/tulu-13B-fp16)
## Prompt template
The following template should be used:
```
<|user|>
prompt goes here
<|assistant|>
```
**Note**: There should be a newline after `<|assistant|>`. This appears to be very important for getting this model to respond correctly.
In other words, the prompt is:
```
<|user|>\nprompt goes here\n<|assistant|>\n
```
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/Jq4vkcDakD)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: Oscar Rangel, Eugene Pentland, Talal Aujan, Cory Kujawski, Luke, Asp the Wyvern, Ai Maven, Pyrater, Alps Aficionado, senxiiz, Willem Michiel, Junyu Yang, trip7s trip, Sebastain Graf, Joseph William Delisle, Lone Striker, Jonathan Leane, Johann-Peter Hartmann, David Flickinger, Spiking Neurons AB, Kevin Schuppel, Mano Prime, Dmitriy Samsonov, Sean Connelly, Nathan LeClaire, Alain Rossmann, Fen Risland, Derek Yates, Luke Pendergrass, Nikolai Manek, Khalefa Al-Ahmad, Artur Olbinski, John Detwiler, Ajan Kanaga, Imad Khwaja, Trenton Dambrowitz, Kalila, vamX, webtim, Illia Dulskyi.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Allen AI's Tulu 13B
# Tulu 13B
This model is a 13B LLaMa model finetuned on a mixture of instruction datasets (FLAN V2, CoT, Dolly, Open Assistant 1, GPT4-Alpaca, Code-Alpaca, and ShareGPT).
*Please note this is a model diff - see below for usage instructions*.
This was trained as part of the paper [How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources](https://arxiv.org/abs/2306.04751).
The codebase used to train and evaluate this model can be found at [https://github.com/allenai/open-instruct](https://github.com/allenai/open-instruct).
This model is licensed under the AI model license given in LICENSE.txt along with the original Llama license (llama_license.txt).
## Usage
We assume you have access to a LLaMa model in HF format already. You can find details on getting access and converting the model here:
[https://huggingface.co/docs/transformers/main/model_doc/llama](https://huggingface.co/docs/transformers/main/model_doc/llama)
Clone [https://github.com/allenai/open-instruct](https://github.com/allenai/open-instruct) and install the required dependencies, or just copy `scripts/weight_diff.py`
and install the minimal requirements listed in `weight-diff-requirements.txt`. Then download or clone this model diff to the same machine.
Then, run:
```bash
python scripts/weight_diff.py recover --path_raw ${hf_llama_path} --path_tuned ${output_path} --path_diff ${diff_location}
```
And you will have a recovered model! Note this takes up a decent amount of RAM, especially for the larger models.
## Input Format
The model is trained to use the following format (note the newlines):
```
<|user|>
Your message here!
<|assistant|>
```
For best results, format all inputs in this manner.
## Performance
Here is the performance of this model across benchmarks explored in our paper [How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources](https://arxiv.org/abs/2306.04751):
| MMLU 0-shot | MMLU 5-shot | GSM Direct | GSM CoT | BBH Direct | BBH CoT | TydiQA Gold-Passage | TydiQA Closed-book | Codex-Eval Pass@1 | Codex-Eval Pass@10 | AlpacaFarm vs Davinci-003 | Average |
|:-----------:|:-----------:|:----------:|:-------:|:----------:|:-------:|:-------------------:|:------------------:|:-----------------:|:------------------:|:-------------------------:|---------|
| 49.2 | 51.8 | 5.0 | 36.5 | 41.3 | 42.8 | 46.1 | 9.2 | 21.3 | 35.0 | 53.9 |37.2 |
If you use this model, please cite our work, the llama paper, and the original datasets:
```
@misc{wang2023far,
title={How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources},
author={Yizhong Wang and Hamish Ivison and Pradeep Dasigi and Jack Hessel and Tushar Khot and Khyathi Raghavi Chandu and David Wadden and Kelsey MacMillan and Noah A. Smith and Iz Beltagy and Hannaneh Hajishirzi},
year={2023},
eprint={2306.04751},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
```
@misc{touvron2023llama,
title={LLaMA: Open and Efficient Foundation Language Models},
author={Hugo Touvron and Thibaut Lavril and Gautier Izacard and Xavier Martinet and Marie-Anne Lachaux and Timothée Lacroix and Baptiste Rozière and Naman Goyal and Eric Hambro and Faisal Azhar and Aurelien Rodriguez and Armand Joulin and Edouard Grave and Guillaume Lample},
year={2023},
eprint={2302.13971},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
```
@misc{dolly,
author = {Databricks},
title = {Free Dolly: Introducing the World's First Truly Open Instruction-Tuned LLM},
year = {2023},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {Blog post},
url = {https://www.databricks.com/blog/2023/04/12/dolly-first-open-commercially-viable-instruction-tuned-llm}
}
```
```
@article{longpre2023flan,
title={The Flan Collection: Designing Data and Methods for Effective Instruction Tuning},
author={Longpre, Shayne and Hou, Le and Vu, Tu and Webson, Albert and Chung, Hyung Won and Tay, Yi and Zhou, Denny and Le, Quoc V and Zoph, Barret and Wei, Jason and others},
journal={arXiv preprint arXiv:2301.13688},
year={2023}
}
```
```
@misc{köpf2023openassistant,
title={OpenAssistant Conversations -- Democratizing Large Language Model Alignment},
author={Andreas Köpf and Yannic Kilcher and Dimitri von Rütte and Sotiris Anagnostidis and Zhi-Rui Tam and Keith Stevens and Abdullah Barhoum and Nguyen Minh Duc and Oliver Stanley and Richárd Nagyfi and Shahul ES and Sameer Suri and David Glushkov and Arnav Dantuluri and Andrew Maguire and Christoph Schuhmann and Huu Nguyen and Alexander Mattick},
year={2023},
eprint={2304.07327},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
```
@article{peng2023instruction,
title={Instruction Tuning with GPT-4},
author={Peng, Baolin and Li, Chunyuan and He, Pengcheng and Galley, Michel and Gao, Jianfeng},
journal={arXiv preprint arXiv:2304.03277},
year={2023}
}
```
```
@misc{codealpaca,
author = {Sahil Chaudhary},
title = {Code Alpaca: An Instruction-following LLaMA model for code generation},
year = {2023},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/sahil280114/codealpaca}},
}
```
|
TheBloke/Manticore-13B-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:48Z | 9 | 1 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-27T00:45:15Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# OpenAccess AI Collective's Manticore 13B fp16
This is fp16 pytorch format model files for [OpenAccess AI Collective's Manticore 13B](https://huggingface.co/openaccess-ai-collective/manticore-13b) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 13b LoRA](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Manticore-13B-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/Manticore-13B-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Manticore-13B-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/openaccess-ai-collective/manticore-13b)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/Manticore-13B-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: OpenAccess AI Collective's Manticore 13B
# Manticore 13B - (previously Wizard Mega)
**[💵 Donate to OpenAccess AI Collective](https://github.com/sponsors/OpenAccess-AI-Collective) to help us keep building great tools and models!**
Questions, comments, feedback, looking to donate, or want to help? Reach out on our [Discord](https://discord.gg/EqrvvehG) or email [[email protected]](mailto:[email protected])
Manticore 13B is a Llama 13B model fine-tuned on the following datasets:
- [ShareGPT](https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered) - based on a cleaned and de-suped subset
- [WizardLM](https://huggingface.co/datasets/ehartford/WizardLM_alpaca_evol_instruct_70k_unfiltered)
- [Wizard-Vicuna](https://huggingface.co/datasets/ehartford/wizard_vicuna_70k_unfiltered)
- [subset of QingyiSi/Alpaca-CoT for roleplay and CoT](https://huggingface.co/QingyiSi/Alpaca-CoT)
- [GPT4-LLM-Cleaned](https://huggingface.co/datasets/teknium/GPT4-LLM-Cleaned)
- [GPTeacher-General-Instruct](https://huggingface.co/datasets/teknium/GPTeacher-General-Instruct)
- ARC-Easy & ARC-Challenge - instruct augmented for detailed responses
- mmlu: instruct augmented for detailed responses subset including
- abstract_algebra
- conceptual_physics
- formal_logic
- high_school_physics
- logical_fallacies
- [hellaswag](https://huggingface.co/datasets/hellaswag) - 5K row subset of instruct augmented for concise responses
- [metaeval/ScienceQA_text_only](https://huggingface.co/datasets/metaeval/ScienceQA_text_only) - instruct for concise responses
- [openai/summarize_from_feedback](https://huggingface.co/datasets/openai/summarize_from_feedback) - instruct augmented tl;dr summarization
# Demo
Try out the model in HF Spaces. The demo uses a quantized GGML version of the model to quickly return predictions on smaller GPUs (and even CPUs). Quantized GGML may have some minimal loss of model quality.
- https://huggingface.co/spaces/openaccess-ai-collective/manticore-ggml
## Release Notes
- https://wandb.ai/wing-lian/manticore-13b/runs/nq3u3uoh/workspace
## Build
Manticore was built with [Axolotl](https://github.com/OpenAccess-AI-Collective/axolotl) on 8xA100 80GB
- Preview Release: 3 epochs taking approximately 24 hours.
- The configuration to duplicate this build is provided in this repo's [/config folder](https://huggingface.co/openaccess-ai-collective/manticore-13b/tree/main/configs).
## Bias, Risks, and Limitations
Manticore has not been aligned to human preferences with techniques like RLHF or deployed with in-the-loop filtering of responses like ChatGPT, so the model can produce problematic outputs (especially when prompted to do so).
Manticore was fine-tuned from the base model LlaMa 13B, please refer to its model card's Limitations Section for relevant information.
## Examples
````
### Instruction: write Python code that returns the first n numbers of the Fibonacci sequence using memoization.
### Assistant:
Here’s an example of how you can use memoization to efficiently calculate the first n numbers of the Fibonacci sequence in Python:
```
def fib(n):
# Memoize the last calculated Fibonacci number for each index
cache = {}
def fib_memoized(n):
if n not in cache:
cache[n] = [0, 1]
else:
return cache[n][-1], cache[n][-2]
# Recursive helper function to calculate the Fibonacci sequence up to a given index
def fib_helper(n):
if n == 0 or n == 1:
return 0, 1
else:
return fib_helper(n - 1) + fib_helper(n - 2)
# Calculate the first n numbers of the Fibonacci sequence using memoization
def fib_memoized_seq(n):
if n == 0 or n == 1:
return 0, 1
else:
result = fib_helper(n - 1) + fib_helper(n - 2)
cache[n] = [result]
return result, cache[n][-1], cache[n][-2]
# Example usage
for i in range(len(fib_memoized_seq(3))):
print(fib_memoized_seq(3)[i])
```
In this implementation, we use a cache dictionary to memoize the last calculated Fibonacci number for each index. We define two helper functions: fib_helper, which recursively calculates the Fibonacci sequence up to a given index, and fib_memoized, which uses memoization to efficiently calculate the first n numbers of the Fibonacci sequence. Finally, we define the fib_memoized_seq function, which is our main function for calculating the first n numbers of the Fibonacci sequence using memoization.
````
```
USER: Finish the joke, a mechanic and a car salesman walk into a bar...
ASSISTANT: … and when they get up to leave, they realize that they’ve left their keys in their cars. The mechanic says “I know how to fix this” and opens his toolbox to grab something. Then he stops and looks at the salesman. He says “You drive a Toyota, right?” The salesman nods. The mechanic then takes out a tiny screwdriver from his toolbox and starts working on the lock of the car door. The salesman is amazed and asks how he knew that. The mechanic replies “I can tell by the way you parked your car.”
```
```
USER: Describe how a tsunami would affect the residents of Denver Colorado
ASSISTANT: The impact of a tsunami in Denver, Colorado would likely be minimal. As there are no coastal areas or bodies of water nearby, a tsunami generated from an earthquake or volcanic activity elsewhere in the world would have little effect on the city and its residents. Even if a large tsunami were to hit the coastline of California or Mexico, the distance between Denver and these locations means that any impacts on the city itself would be minimal.
```
```
USER: If you could do one thing right now, what would it be?
ASSISTANT: I’d hug my loved ones a little tighter today and tell them how much they mean to me.
USER: And who are your loved ones?
ASSISTANT: My closest friends and family members.
USER: What are their names?
ASSISTANT: My two best friends are named Emily and Jessica, my mom is Mary, and my dad is David. Answering the question with a bit more information about who they are or what they mean to me would help provide a more personalized response.
```
|
TheBloke/Minotaur-13B-fixed-SuperHOT-8K-fp16 | TheBloke | 2023-07-09T20:24:48Z | 12 | 2 | transformers | [
"transformers",
"pytorch",
"llama",
"text-generation",
"custom_code",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | text-generation | 2023-06-27T01:22:02Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# OpenAccess AI Collective's Minotaur 13B Fixed fp16
This is fp16 pytorch format model files for [OpenAccess AI Collective's Minotaur 13B Fixed](https://huggingface.co/openaccess-ai-collective/minotaur-13b-fixed) merged with [Kaio Ken's SuperHOT 8K](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test).
[Kaio Ken's SuperHOT 13b LoRA](https://huggingface.co/kaiokendev/superhot-13b-8k-no-rlhf-test) is merged on to the base model, and then 8K context can be achieved during inference by using `trust_remote_code=True`.
Note that `config.json` has been set to a sequence length of 8192. This can be modified to 4096 if you want to try with a smaller sequence length.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/Minotaur-13B-fixed-SuperHOT-8K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU inference](https://huggingface.co/TheBloke/Minotaur-13B-fixed-SuperHOT-8K-GGML)
* [Unquantised SuperHOT fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Minotaur-13B-fixed-SuperHOT-8K-fp16)
* [Unquantised base fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/openaccess-ai-collective/minotaur-13b-fixed)
## How to use this model from Python code
First make sure you have Einops installed:
```
pip3 install auto-gptq
```
Then run the following code. `config.json` has been default to a sequence length of 8192, but you can also configure this in your Python code.
The provided modelling code, activated with `trust_remote_code=True` will automatically set the `scale` parameter from the configured `max_position_embeddings`. Eg for 8192, `scale` is set to `4`.
```python
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, pipeline
import argparse
model_name_or_path = "TheBloke/Minotaur-13B-fixed-SuperHOT-8K-fp16"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
config = AutoConfig.from_pretrained(model_name_or_path, trust_remote_code=True)
# Change this to the sequence length you want
config.max_position_embeddings = 8192
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
config=config,
trust_remote_code=True,
device_map='auto')
# Note: check to confirm if this is correct prompt template is correct for this model!
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Using other UIs: monkey patch
Provided in the repo is `llama_rope_scaled_monkey_patch.py`, written by @kaiokendev.
It can be theoretically be added to any Python UI or custom code to enable the same result as `trust_remote_code=True`. I have not tested this, and it should be superseded by using `trust_remote_code=True`, but I include it for completeness and for interest.
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
**Patreon special mentions**: zynix , ya boyyy, Trenton Dambrowitz, Imad Khwaja, Alps Aficionado, chris gileta, John Detwiler, Willem Michiel, RoA, Mano Prime, Rainer Wilmers, Fred von Graf, Matthew Berman, Ghost , Nathan LeClaire, Iucharbius , Ai Maven, Illia Dulskyi, Joseph William Delisle, Space Cruiser, Lone Striker, Karl Bernard, Eugene Pentland, Greatston Gnanesh, Jonathan Leane, Randy H, Pierre Kircher, Willian Hasse, Stephen Murray, Alex , terasurfer , Edmond Seymore, Oscar Rangel, Luke Pendergrass, Asp the Wyvern, Junyu Yang, David Flickinger, Luke, Spiking Neurons AB, subjectnull, Pyrater, Nikolai Manek, senxiiz, Ajan Kanaga, Johann-Peter Hartmann, Artur Olbinski, Kevin Schuppel, Derek Yates, Kalila, K, Talal Aujan, Khalefa Al-Ahmad, Gabriel Puliatti, John Villwock, WelcomeToTheClub, Daniel P. Andersen, Preetika Verma, Deep Realms, Fen Risland, trip7s trip, webtim, Sean Connelly, Michael Levine, Chris McCloskey, biorpg, vamX, Viktor Bowallius, Cory Kujawski.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: Kaio Ken's SuperHOT 8K
### SuperHOT Prototype 2 w/ 8K Context
This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in [the github blog](https://kaiokendev.github.io/til#extending-context-to-8k).
Tests have shown that the model does indeed leverage the extended context at 8K.
You will need to **use either the monkeypatch** or, if you are already using the monkeypatch, **change the scaling factor to 0.25 and the maximum sequence length to 8192**
#### Looking for Merged & Quantized Models?
- 30B 4-bit CUDA: [tmpupload/superhot-30b-8k-4bit-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-safetensors)
- 30B 4-bit CUDA 128g: [tmpupload/superhot-30b-8k-4bit-128g-safetensors](https://huggingface.co/tmpupload/superhot-30b-8k-4bit-128g-safetensors)
#### Training Details
I trained the LoRA with the following configuration:
- 1200 samples (~400 samples over 2048 sequence length)
- learning rate of 3e-4
- 3 epochs
- The exported modules are:
- q_proj
- k_proj
- v_proj
- o_proj
- no bias
- Rank = 4
- Alpha = 8
- no dropout
- weight decay of 0.1
- AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
- Trained on 4-bit base model
# Original model card: OpenAccess AI Collective's Minotaur 13B Fixed
[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
**[💵 Donate to OpenAccess AI Collective](https://github.com/sponsors/OpenAccess-AI-Collective) to help us keep building great tools and models!**
# Due to a bug, the initial release of Minotaur 13B dropped a few datasets during training. We have corrected the issue and this is the retrained model
The affected datasets include:
- prose generation
- classification
- coding
# Minotaur 13B (FIXED)
Minotaur 13B is an instruct fine-tuned model on top of LlaMA-13B. Minotaur 13B is fine-tuned **on only completely open datasets** making this model reproducible by anyone.
Questions, comments, feedback, looking to donate, or want to help? Reach out on our [Discord](https://discord.gg/PugNNHAF5r) or email [[email protected]](mailto:[email protected])
# Prompts
Chat only style prompts using `USER:`,`ASSISTANT:`.
<img src="https://huggingface.co/openaccess-ai-collective/minotaur-13b/resolve/main/minotaur.png" alt="minotaur" width="600" height="500"/>
# Training Datasets
Minotaur 13B model is fine-tuned on the following openly available datasets:
- [WizardLM](https://huggingface.co/datasets/ehartford/WizardLM_alpaca_evol_instruct_70k_unfiltered)
- [subset of QingyiSi/Alpaca-CoT for roleplay and CoT](https://huggingface.co/QingyiSi/Alpaca-CoT)
- [GPTeacher-General-Instruct](https://huggingface.co/datasets/teknium/GPTeacher-General-Instruct)
- [metaeval/ScienceQA_text_only](https://huggingface.co/datasets/metaeval/ScienceQA_text_only) - instruct for concise responses
- [openai/summarize_from_feedback](https://huggingface.co/datasets/openai/summarize_from_feedback) - instruct augmented tl;dr summarization
- [camel-ai/math](https://huggingface.co/datasets/camel-ai/math)
- [camel-ai/physics](https://huggingface.co/datasets/camel-ai/physics)
- [camel-ai/chemistry](https://huggingface.co/datasets/camel-ai/chemistry)
- [camel-ai/biology](https://huggingface.co/datasets/camel-ai/biology)
- [winglian/evals](https://huggingface.co/datasets/winglian/evals) - instruct augmented datasets
- custom sysnthetic datasets around misconceptions, in-context qa, jokes, N-tasks problems, and context-insensitivity
- ARC-Easy & ARC-Challenge - instruct augmented for detailed responses, derived from the `train` split
- [hellaswag](https://huggingface.co/datasets/hellaswag) - 30K+ rows of instruct augmented for detailed explanations w 30K+ rows, derived from the `train` split
- [riddle_sense](https://huggingface.co/datasets/riddle_sense) - instruct augmented, derived from the `train` split
- [gsm8k](https://huggingface.co/datasets/gsm8k) - instruct augmented, derived from the `train` split
- prose generation
# Shoutouts
Special thanks to Nanobit for helping with Axolotl and TheBloke for quantizing these models are more accessible to all.
# Demo
HF Demo in Spaces available in the [Community ChatBot Arena](https://huggingface.co/spaces/openaccess-ai-collective/rlhf-arena) under the OAAIC Chatbots tab.
## Release Notes
- https://wandb.ai/wing-lian/minotaur-13b/runs/5ystr7w6/workspace
## Build
Minotaur was built with [Axolotl](https://github.com/OpenAccess-AI-Collective/axolotl) on 6XA100 80GB
- 1 epochs taking approximately 7.5 hours
## Bias, Risks, and Limitations
Minotaur has not been aligned to human preferences with techniques like RLHF or deployed with in-the-loop filtering of responses like ChatGPT, so the model can produce problematic outputs (especially when prompted to do so).
Minotaur was fine-tuned from the base model LLaMA-13B, please refer to its model card's Limitations Section for relevant information. (included below)
## Benchmarks
hf-causal-experimental (pretrained=openaccess-ai-collective/minotaur-13b-fixed), limit: None, provide_description: False, num_fewshot: 0, batch_size: None
| Task |Version| Metric |Value | |Stderr|
|-------------|------:|--------|-----:|---|-----:|
|arc_challenge| 0|acc |0.4778|_ |0.0146|
| | |acc_norm|0.4932|_ |0.0146|
|arc_easy | 0|acc |0.7786|_ |0.0085|
| | |acc_norm|0.7487|_ |0.0089|
|boolq | 1|acc |0.8141|_ |0.0068|
|hellaswag | 0|acc |0.6072|_ |0.0049|
| | |acc_norm|0.7929|_ |0.0040|
|openbookqa | 0|acc |0.3420|_ |0.0212|
| | |acc_norm|0.4580|_ |0.0223|
|piqa | 0|acc |0.8025|_ |0.0093|
| | |acc_norm|0.8107|_ |0.0091|
|winogrande | 0|acc |0.7285|_ |0.0125|
## Examples - results may vary based on temperature (0.7 for this) and other settings
```
The following is a chat between a USER and a friendly and helpful ASSISTANT.
USER: Solve for x: 5x = 3x + 10
ASSISTANT:A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
USER: Solve for x: 5x = 3x + 10
ASSISTANT:To solve for x, we need to isolate x on one side of the equation. We can do this by following these steps:
1. Subtract 3x from both sides of the equation:
5x - 3x = 3x + 10 - 3x
2x = 10
2. Divide both sides of the equation by 2:
2x/2 = 10/2
x = 5
```
```
The following is a chat between a USER and a friendly and helpful ASSISTANT.
USER: You have 3 apples. You eat 2 pears. How many apples do you have left?
ASSISTANT:You still have 3 apples. Eating pears does not affect the number of apples you have.
```
```
The following is a chat between a USER and a friendly and helpful ASSISTANT.
USER: List three words that start with "ex". What is 33 + 77?
ASSISTANT:Three words that start with "ex" are "express", "exact", and "excite". 33 + 77 = 110.
```
```
The following is a chat between a USER and a friendly and helpful ASSISTANT.
USER: Write a haiku about dinosaurs.
ASSISTANT:Titanic creatures roam,
Thunderous footsteps echo,
Extinct but not forgotten.
```
meh, 6/7/7 rather than 5/7/5
```
A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
USER: Create an original short story in the style of J.K Rowling about the Weesley family going on a vacation in Australia whith Harry Potter and Hermione Grainger tagging along.
ASSISTANT:Once upon a time, there was a magical family named the Weesleys. They lived in a cozy little cottage at the edge of Hogsmeade village, where they ran a charming bed and breakfast called "The Flying Ford Anglia." A
s any other magical family, they enjoyed adventures that brought them close to nature, so they decided to go on a trip to Australia. They invited their dear friends Harry Potter and Hermione Grainger to join them on this ex
citing journey.
The Weesleys had been planning for months and had everything ready for their trip. Their children, Freddie and Roxanne were thrilled to visit the land down under, and their parents, George and Angelina, could not wait to ex
plore the beautiful country with their friends. They would be leaving soon, but before departing, there was one small problem - the luggage.
"We can't possibly take all these trunks!" exclaimed Freddie as he stared at the mountain of bags stacked up in front of him.
His father looked at him seriously. "You see, son," he began, "our luggage is crucial because we have to pack all our magical essentials like wand holsters, extra wand chargers, spell books, potions ingredients, and much mo
re."
Roxanne interrupted, "But daddy, we don't even use magic in everyday life! And what if it rains? We'll look weird carrying umbrellas around!"
Her mother joined in. "My dear, you know how unpredictable the weather can be. We must always be prepared for unexpected circumstances. Don't forget that even Muggles carry umbrellas."
After hours of debating, they finally managed to narrow down their luggage list and fit all the necessary supplies into several large trunks. The day had come; they were ready to leave for their grand adventure!
As the Weesleys boarded the Portkey that would transport them across the world, their wands began to glow softly, indicating that they had enough energy to make the journey. The Portkey dropped them off in Sydney, right in
front of the magnificent Opera House.
They spent the first few days exploring the city, marveling at the iconic architecture and tasting local cuisine. Then, as planned, they headed north to visit the Great Barrier Reef, one of the most famous natural wonders o
f the world.
Harry and Hermione joined them during this leg of the trip, which made it even more enjoyable. Harry regaled them with tales of his own travels while Hermione shared her extensive knowledge of plants, animals, and the envir
onment.
Soon, they arrived at a quaint town nestled among vibrant green hills and surrounded by vast cattle farms. It was here that they would learn about Aboriginal culture and see some truly unique sights.
One morning, after enjoying a hearty breakfast, they set out to explore the local area. They visited a nearby art gallery that showcased amazing Indigenous works of art, including traditional paintings, sculptures, and text
iles. Afterward, they attended a didgeridoo concert given by a talented young musician who captivated everyone with his soulful tunes.
The following day, they embarked on a bushwalk through the rainforest trails. The air was cool and fresh, and the towering trees seemed to reach for the sky. Hermione took great pleasure in identifying various species of pl
ants and explaining their medicinal properties.
As they walked deeper into the forest, they heard faint chattering noises coming from high above. To their surprise, they spotted a group of small, furry creatures perched on a tree branch. They turned out to be native Aust
ralian possums, which delighted the children, especially Freddie, who could communicate with them using sign language.
Later that afternoon, they visited Uluru, also known as Ayers Rock, a massive sandstone monolith located in the heart of the Outback. The sight of its magnificent red hues against the blue sky left them all speechless. They
spent the rest of the day exploring the rock formation, taking in the breathtaking views and learning about its significance to the Anangu people.
Their last destination was Queensland, home to the world-famous Great Barrier Reef. They spent two days snorkeling, diving, and swimming among the colorful coral reefs, watching schools of fish, sea turtles, and dolphins sw
im past them.
As the sun began to set over the crystal-clear waters, Harry spoke up. "I think it's time we return to reality, don't you agree?" he said, looking at his wife. "This trip has been fantastic, but we can't stay forever."
Everyone nodded in agreement as they prepared themselves for their final journey back to Sydney. When they got there, they gathered on the balcony of their hotel room, reminiscing about the incredible memories they had crea
ted together.
"Thank you for inviting us along on your adventure, Weesleys," said Harry. "It was an honor and a privilege to share this experience with such kind and generous friends. I hope we can do it again sometime."
With tears in their eyes and heavy hearts, they bid each other farewell, promising to keep in touch and meet again soon. And so, their epic journey came to an end. But the memories would remain with them forever, reminding
them of the power of friendship, the beauty of nature, and the importance of discovering new worlds.
```
|
dp15/poca-SoccerTwosv2 | dp15 | 2023-07-09T19:50:02Z | 1 | 0 | ml-agents | [
"ml-agents",
"tensorboard",
"onnx",
"SoccerTwos",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-SoccerTwos",
"region:us"
] | reinforcement-learning | 2023-07-09T19:45:35Z | ---
library_name: ml-agents
tags:
- SoccerTwos
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-SoccerTwos
---
# **poca** Agent playing **SoccerTwos**
This is a trained model of a **poca** agent playing **SoccerTwos**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: dp15/poca-SoccerTwosv2
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
TalesLF/a2c-PandaReachDense-v2 | TalesLF | 2023-07-09T19:40:12Z | 0 | 0 | stable-baselines3 | [
"stable-baselines3",
"PandaReachDense-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] | reinforcement-learning | 2023-07-09T19:36:52Z | ---
library_name: stable-baselines3
tags:
- PandaReachDense-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: A2C
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: PandaReachDense-v2
type: PandaReachDense-v2
metrics:
- type: mean_reward
value: -0.66 +/- 0.21
name: mean_reward
verified: false
---
# **A2C** Agent playing **PandaReachDense-v2**
This is a trained model of a **A2C** agent playing **PandaReachDense-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
...
```
|
Jonathaniu/alpaca-bitcoin-tweets-sentiment | Jonathaniu | 2023-07-09T19:32:27Z | 5 | 1 | peft | [
"peft",
"region:us"
] | null | 2023-07-08T01:05:11Z | ---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: True
- load_in_4bit: False
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.4.0.dev0
|
RogerB/deberta-base-finetuned-kintweetsE | RogerB | 2023-07-09T19:29:34Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"deberta",
"fill-mask",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-07-09T19:09:39Z | ---
license: mit
tags:
- generated_from_trainer
model-index:
- name: deberta-base-finetuned-kintweetsE
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. -->
# deberta-base-finetuned-kintweetsE
This model is a fine-tuned version of [microsoft/deberta-base](https://huggingface.co/microsoft/deberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.4010
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 4.7396 | 1.0 | 1000 | 3.9160 |
| 3.7652 | 2.0 | 2000 | 3.4778 |
| 3.5318 | 3.0 | 3000 | 3.3727 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
dp15/poca-SoccerTwos | dp15 | 2023-07-09T19:28:19Z | 1 | 0 | ml-agents | [
"ml-agents",
"tensorboard",
"onnx",
"SoccerTwos",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-SoccerTwos",
"region:us"
] | reinforcement-learning | 2023-07-09T17:03:18Z | ---
library_name: ml-agents
tags:
- SoccerTwos
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-SoccerTwos
---
# **poca** Agent playing **SoccerTwos**
This is a trained model of a **poca** agent playing **SoccerTwos**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: dp15/poca-SoccerTwos
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
RogerB/xlm-roberta-base-finetuned-kintweetsE | RogerB | 2023-07-09T19:08:26Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"fill-mask",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-07-09T18:36:00Z | ---
license: mit
tags:
- generated_from_trainer
model-index:
- name: xlm-roberta-base-finetuned-kintweetsE
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. -->
# xlm-roberta-base-finetuned-kintweetsE
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.9498
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.7175 | 1.0 | 1000 | 3.1645 |
| 3.2319 | 2.0 | 2000 | 3.0430 |
| 3.0974 | 3.0 | 3000 | 3.0093 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
JamesVorder/PPO-LunarLander-v2 | JamesVorder | 2023-07-09T19:04:27Z | 0 | 0 | stable-baselines3 | [
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] | reinforcement-learning | 2023-07-09T19:04:09Z | ---
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: 254.45 +/- 24.22
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
...
```
|
Bisht0538/gauravbisht | Bisht0538 | 2023-07-09T18:55:53Z | 183 | 0 | transformers | [
"transformers",
"pytorch",
"tf",
"jax",
"rust",
"bart",
"text2text-generation",
"summarization",
"en",
"dataset:cnn_dailymail",
"arxiv:1910.13461",
"license:mit",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | summarization | 2023-07-09T17:31:09Z | ---
language:
- en
tags:
- summarization
license: mit
thumbnail: https://huggingface.co/front/thumbnails/facebook.png
datasets:
- cnn_dailymail
model-index:
- name: facebook/bart-large-cnn
results:
- task:
type: summarization
name: Summarization
dataset:
name: cnn_dailymail
type: cnn_dailymail
config: 3.0.0
split: train
metrics:
- name: ROUGE-1
type: rouge
value: 42.9486
verified: true
- name: ROUGE-2
type: rouge
value: 20.8149
verified: true
- name: ROUGE-L
type: rouge
value: 30.6186
verified: true
- name: ROUGE-LSUM
type: rouge
value: 40.0376
verified: true
- name: loss
type: loss
value: 2.529000997543335
verified: true
- name: gen_len
type: gen_len
value: 78.5866
verified: true
---
# BART (large-sized model), fine-tuned on CNN Daily Mail
BART model pre-trained on English language, and fine-tuned on [CNN Daily Mail](https://huggingface.co/datasets/cnn_dailymail). It was introduced in the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) by Lewis et al. and first released in [this repository (https://github.com/pytorch/fairseq/tree/master/examples/bart).
Disclaimer: The team releasing BART did not write a model card for this model so this model card has been written by the Hugging Face team.
## Model description
BART is a transformer encoder-encoder (seq2seq) model with a bidirectional (BERT-like) encoder and an autoregressive (GPT-like) decoder. BART is pre-trained by (1) corrupting text with an arbitrary noising function, and (2) learning a model to reconstruct the original text.
BART is particularly effective when fine-tuned for text generation (e.g. summarization, translation) but also works well for comprehension tasks (e.g. text classification, question answering). This particular checkpoint has been fine-tuned on CNN Daily Mail, a large collection of text-summary pairs.
## Intended uses & limitations
You can use this model for text summarization.
### How to use
Here is how to use this model with the [pipeline API](https://huggingface.co/transformers/main_classes/pipelines.html):
```python
from transformers import pipeline
summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
ARTICLE = """ New York (CNN)When Liana Barrientos was 23 years old, she got married in Westchester County, New York.
A year later, she got married again in Westchester County, but to a different man and without divorcing her first husband.
Only 18 days after that marriage, she got hitched yet again. Then, Barrientos declared "I do" five more times, sometimes only within two weeks of each other.
In 2010, she married once more, this time in the Bronx. In an application for a marriage license, she stated it was her "first and only" marriage.
Barrientos, now 39, is facing two criminal counts of "offering a false instrument for filing in the first degree," referring to her false statements on the
2010 marriage license application, according to court documents.
Prosecutors said the marriages were part of an immigration scam.
On Friday, she pleaded not guilty at State Supreme Court in the Bronx, according to her attorney, Christopher Wright, who declined to comment further.
After leaving court, Barrientos was arrested and charged with theft of service and criminal trespass for allegedly sneaking into the New York subway through an emergency exit, said Detective
Annette Markowski, a police spokeswoman. In total, Barrientos has been married 10 times, with nine of her marriages occurring between 1999 and 2002.
All occurred either in Westchester County, Long Island, New Jersey or the Bronx. She is believed to still be married to four men, and at one time, she was married to eight men at once, prosecutors say.
Prosecutors said the immigration scam involved some of her husbands, who filed for permanent residence status shortly after the marriages.
Any divorces happened only after such filings were approved. It was unclear whether any of the men will be prosecuted.
The case was referred to the Bronx District Attorney\'s Office by Immigration and Customs Enforcement and the Department of Homeland Security\'s
Investigation Division. Seven of the men are from so-called "red-flagged" countries, including Egypt, Turkey, Georgia, Pakistan and Mali.
Her eighth husband, Rashid Rajput, was deported in 2006 to his native Pakistan after an investigation by the Joint Terrorism Task Force.
If convicted, Barrientos faces up to four years in prison. Her next court appearance is scheduled for May 18.
"""
print(summarizer(ARTICLE, max_length=130, min_length=30, do_sample=False))
>>> [{'summary_text': 'Liana Barrientos, 39, is charged with two counts of "offering a false instrument for filing in the first degree" In total, she has been married 10 times, with nine of her marriages occurring between 1999 and 2002. She is believed to still be married to four men.'}]
```
### BibTeX entry and citation info
```bibtex
@article{DBLP:journals/corr/abs-1910-13461,
author = {Mike Lewis and
Yinhan Liu and
Naman Goyal and
Marjan Ghazvininejad and
Abdelrahman Mohamed and
Omer Levy and
Veselin Stoyanov and
Luke Zettlemoyer},
title = {{BART:} Denoising Sequence-to-Sequence Pre-training for Natural Language
Generation, Translation, and Comprehension},
journal = {CoRR},
volume = {abs/1910.13461},
year = {2019},
url = {http://arxiv.org/abs/1910.13461},
eprinttype = {arXiv},
eprint = {1910.13461},
timestamp = {Thu, 31 Oct 2019 14:02:26 +0100},
biburl = {https://dblp.org/rec/journals/corr/abs-1910-13461.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
} |
aryjessen/SkyHawk | aryjessen | 2023-07-09T18:54:28Z | 0 | 0 | null | [
"text-to-image",
"en",
"region:us"
] | text-to-image | 2023-07-09T18:49:43Z | ---
language:
- en
pipeline_tag: text-to-image
--- |
nacicansa/bertin-gpt-j-6B-alpaca-adapter-with-openassistant-oasst1 | nacicansa | 2023-07-09T18:48:05Z | 2 | 0 | peft | [
"peft",
"region:us"
] | null | 2023-07-09T12:45:34Z | ---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: True
- load_in_4bit: False
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: True
- load_in_4bit: False
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: True
- load_in_4bit: False
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: True
- load_in_4bit: False
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: True
- load_in_4bit: False
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: True
- load_in_4bit: False
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.4.0.dev0
- PEFT 0.4.0.dev0
- PEFT 0.4.0.dev0
- PEFT 0.4.0.dev0
- PEFT 0.4.0.dev0
- PEFT 0.4.0.dev0
|
lucasbertola/Reinforce-CartPole | lucasbertola | 2023-07-09T18:36:22Z | 0 | 0 | null | [
"CartPole-v1",
"reinforce",
"reinforcement-learning",
"custom-implementation",
"deep-rl-class",
"model-index",
"region:us"
] | reinforcement-learning | 2023-07-09T15:29:01Z | ---
tags:
- CartPole-v1
- reinforce
- reinforcement-learning
- custom-implementation
- deep-rl-class
model-index:
- name: Reinforce-CartPole
results:
- metrics:
- type: mean_reward
value: 494.83 +/- 25.41
name: mean_reward
task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: CartPole-v1
type: CartPole-v1
---
# **Reinforce** Agent playing **CartPole-v1**
This is a trained model of a **Reinforce** agent playing **CartPole-v1** .
To learn to use this model and train yours check Unit 4 of the Deep Reinforcement Learning Course: https://huggingface.co/deep-rl-course/unit4/introduction
|
RogerB/roberta-base-finetuned-kintweetsE | RogerB | 2023-07-09T18:28:41Z | 131 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"roberta",
"fill-mask",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-07-09T18:13:57Z | ---
license: mit
tags:
- generated_from_trainer
model-index:
- name: roberta-base-finetuned-kintweetsE
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-base-finetuned-kintweetsE
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.6524
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.1657 | 1.0 | 1000 | 2.8429 |
| 2.8541 | 2.0 | 2000 | 2.6654 |
| 2.7484 | 3.0 | 3000 | 2.6122 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
aphi/Reinforce-CartPole-v1 | aphi | 2023-07-09T18:24:36Z | 0 | 0 | null | [
"CartPole-v1",
"reinforce",
"reinforcement-learning",
"custom-implementation",
"deep-rl-class",
"model-index",
"region:us"
] | reinforcement-learning | 2023-07-09T18:24:27Z | ---
tags:
- CartPole-v1
- reinforce
- reinforcement-learning
- custom-implementation
- deep-rl-class
model-index:
- name: Reinforce-CartPole-v1
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: CartPole-v1
type: CartPole-v1
metrics:
- type: mean_reward
value: 494.30 +/- 17.10
name: mean_reward
verified: false
---
# **Reinforce** Agent playing **CartPole-v1**
This is a trained model of a **Reinforce** agent playing **CartPole-v1** .
To learn to use this model and train yours check Unit 4 of the Deep Reinforcement Learning Course: https://huggingface.co/deep-rl-course/unit4/introduction
|
nikolai40/iam-trocr | nikolai40 | 2023-07-09T18:18:53Z | 46 | 0 | transformers | [
"transformers",
"pytorch",
"vision-encoder-decoder",
"image-text-to-text",
"endpoints_compatible",
"region:us"
] | image-text-to-text | 2023-06-08T10:57:44Z | ---
metrics:
- trocr
- image-to-text
- CER
---
# TrOCR model fine-tuned on IAM dataset using augmentation (stretching and dilation)
Using [trocr-small-stage1](https://huggingface.co/microsoft/trocr-small-stage1) version |
RogerB/distilbert-base-multilingual-cased-finetuned-kintweetsE | RogerB | 2023-07-09T18:12:36Z | 124 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"fill-mask",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-07-09T17:56:03Z | ---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-multilingual-cased-finetuned-kintweetsE
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. -->
# distilbert-base-multilingual-cased-finetuned-kintweetsE
This model is a fine-tuned version of [distilbert-base-multilingual-cased](https://huggingface.co/distilbert-base-multilingual-cased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1438
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.7676 | 1.0 | 1000 | 3.3390 |
| 3.3493 | 2.0 | 2000 | 3.1638 |
| 3.2122 | 3.0 | 3000 | 3.1040 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
krystaleahr/detr-resnet-50_finetuned_candy | krystaleahr | 2023-07-09T18:00:12Z | 187 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"detr",
"object-detection",
"generated_from_trainer",
"dataset:imagefolder",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | object-detection | 2023-07-09T17:10:07Z | ---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- imagefolder
model-index:
- name: detr-resnet-50_finetuned_candy
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. -->
# detr-resnet-50_finetuned_candy
This model is a fine-tuned version of [facebook/detr-resnet-50](https://huggingface.co/facebook/detr-resnet-50) on the imagefolder 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: 1e-05
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 75
### Training results
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
RogerB/distilbert-base-uncased-finetuned-kintweetsE | RogerB | 2023-07-09T17:55:15Z | 124 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"fill-mask",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-07-09T17:47:31Z | ---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-uncased-finetuned-kintweetsE
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. -->
# distilbert-base-uncased-finetuned-kintweetsE
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1921
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.7016 | 1.0 | 1000 | 3.3955 |
| 3.342 | 2.0 | 2000 | 3.2020 |
| 3.2277 | 3.0 | 3000 | 3.1666 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
RogerB/bert-base-uncased-finetuned-kintweetsE | RogerB | 2023-07-09T17:46:49Z | 107 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"bert",
"fill-mask",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-07-09T17:34:25Z | ---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: bert-base-uncased-finetuned-kintweetsE
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. -->
# bert-base-uncased-finetuned-kintweetsE
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.0631
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.6434 | 1.0 | 1000 | 3.2659 |
| 3.2304 | 2.0 | 2000 | 3.0654 |
| 3.106 | 3.0 | 3000 | 3.0272 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
mrizalf7/t5-small-finetuned-indosum-3 | mrizalf7 | 2023-07-09T17:46:49Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"t5",
"text2text-generation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text2text-generation | 2023-07-09T16:18:16Z | ---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: t5-small-finetuned-indosum-3
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. -->
# t5-small-finetuned-indosum-3
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) 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: 2e-05
- train_batch_size: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
- mixed_precision_training: Native AMP
### Framework versions
- Transformers 4.28.0
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
Hedayat-Abrishami/ppo-Pyramid | Hedayat-Abrishami | 2023-07-09T17:45:06Z | 0 | 0 | ml-agents | [
"ml-agents",
"tensorboard",
"onnx",
"Pyramids",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-Pyramids",
"region:us"
] | reinforcement-learning | 2023-07-09T17:44:34Z | ---
library_name: ml-agents
tags:
- Pyramids
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-Pyramids
---
# **ppo** Agent playing **Pyramids**
This is a trained model of a **ppo** agent playing **Pyramids**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: Hedayat-Abrishami/ppo-Pyramid
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
vnktrmnb/bert-base-multilingual-cased-finetuned-tydiqa | vnktrmnb | 2023-07-09T17:42:01Z | 61 | 0 | transformers | [
"transformers",
"tf",
"tensorboard",
"bert",
"question-answering",
"generated_from_keras_callback",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | question-answering | 2023-07-08T19:08:21Z | ---
license: apache-2.0
tags:
- generated_from_keras_callback
model-index:
- name: vnktrmnb/bert-base-multilingual-cased-finetuned-tydiqa
results: []
---
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# vnktrmnb/bert-base-multilingual-cased-finetuned-tydiqa
This model is a fine-tuned version of [bert-base-multilingual-cased](https://huggingface.co/bert-base-multilingual-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Train Loss: 0.6211
- Train End Logits Accuracy: 0.8146
- Train Start Logits Accuracy: 0.8612
- Validation Loss: 0.4720
- Validation End Logits Accuracy: 0.8544
- Validation Start Logits Accuracy: 0.9103
- Epoch: 1
## 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:
- optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 836, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False}
- training_precision: float32
### Training results
| Train Loss | Train End Logits Accuracy | Train Start Logits Accuracy | Validation Loss | Validation End Logits Accuracy | Validation Start Logits Accuracy | Epoch |
|:----------:|:-------------------------:|:---------------------------:|:---------------:|:------------------------------:|:--------------------------------:|:-----:|
| 0.9948 | 0.7232 | 0.7723 | 0.4799 | 0.8500 | 0.9029 | 0 |
| 0.6211 | 0.8146 | 0.8612 | 0.4720 | 0.8544 | 0.9103 | 1 |
### Framework versions
- Transformers 4.30.2
- TensorFlow 2.12.0
- Datasets 2.13.1
- Tokenizers 0.13.3
|
TheBloke/WizardLM-13B-V1.1-GGML | TheBloke | 2023-07-09T17:27:27Z | 0 | 35 | null | [
"arxiv:2304.12244",
"license:other",
"region:us"
] | null | 2023-07-07T16:25:24Z | ---
inference: false
license: other
---
<!-- header start -->
<div style="width: 100%;">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<!-- header end -->
# WizardLM's WizardLM 13B V1.1 GGML
These files are GGML format model files for [WizardLM's WizardLM 13B V1.1](https://huggingface.co/WizardLM/WizardLM-13B-V1.1).
GGML files are for CPU + GPU inference using [llama.cpp](https://github.com/ggerganov/llama.cpp) and libraries and UIs which support this format, such as:
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui)
* [KoboldCpp](https://github.com/LostRuins/koboldcpp)
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui)
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python)
* [ctransformers](https://github.com/marella/ctransformers)
## Update 9th July 2023: GGML k-quants now available
Thanks to the work of LostRuins/concedo, it is now possible to provide 100% working GGML k-quants for models like this which have a non-standard vocab size (32,001).
k-quants have been uploaded and will work with all llama.cpp clients without any changes required.
## Repositories available
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/WizardLM-13B-V1.1-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/WizardLM-13B-V1.1-GGML)
* [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/WizardLM/WizardLM-13B-V1.1)
## Prompt template: Vicuna
```
A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
USER: PROMPT
ASSISTANT:
```
<!-- compatibility_ggml start -->
## Compatibility
**Note:** due to this model having a non-standard vocab size of 32,001, k-quants are slightly larger than they are for other models of the same size and type.
For example, a 13B q4_K_M will be around 150MB larger. Inference speed should not be noticeably affected, and quality will be the same or higher.
### Original llama.cpp quant methods: `q4_0, q4_1, q5_0, q5_1, q8_0`
These are guaranteed to be compatible with any UIs, tools and libraries released since late May. They may be phased out soon, as they are largely superseded by the new k-quant methods.
### New k-quant methods: `q2_K, q3_K_S, q3_K_M, q3_K_L, q4_K_S, q4_K_M, q5_K_S, q6_K`
These new quantisation methods are compatible with llama.cpp as of June 6th, commit `2d43387`.
They are now also compatible with recent releases of text-generation-webui, KoboldCpp, llama-cpp-python, ctransformers, rustformers and most others. For compatibility with other tools and libraries, please check their documentation.
## Explanation of the new k-quant methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
* GGML_TYPE_Q8_K - "type-0" 8-bit quantization. Only used for quantizing intermediate results. The difference to the existing Q8_0 is that the block size is 256. All 2-6 bit dot products are implemented for this quantization type.
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_ggml end -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| wizardlm-13b-v1.1.ggmlv3.q2_K.bin | q2_K | 2 | 5.67 GB| 8.17 GB | New k-quant method. Uses GGML_TYPE_Q4_K for the attention.vw and feed_forward.w2 tensors, GGML_TYPE_Q2_K for the other tensors. |
| wizardlm-13b-v1.1.ggmlv3.q3_K_L.bin | q3_K_L | 3 | 7.07 GB| 9.57 GB | New k-quant method. Uses GGML_TYPE_Q5_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K |
| wizardlm-13b-v1.1.ggmlv3.q3_K_M.bin | q3_K_M | 3 | 6.46 GB| 8.96 GB | New k-quant method. Uses GGML_TYPE_Q4_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K |
| wizardlm-13b-v1.1.ggmlv3.q3_K_S.bin | q3_K_S | 3 | 5.80 GB| 8.30 GB | New k-quant method. Uses GGML_TYPE_Q3_K for all tensors |
| wizardlm-13b-v1.1.ggmlv3.q4_0.bin | q4_0 | 4 | 7.32 GB| 9.82 GB | Original quant method, 4-bit. |
| wizardlm-13b-v1.1.ggmlv3.q4_1.bin | q4_1 | 4 | 8.14 GB| 10.64 GB | Original quant method, 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
| wizardlm-13b-v1.1.ggmlv3.q4_K_M.bin | q4_K_M | 4 | 7.99 GB| 10.49 GB | New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q4_K |
| wizardlm-13b-v1.1.ggmlv3.q4_K_S.bin | q4_K_S | 4 | 7.49 GB| 9.99 GB | New k-quant method. Uses GGML_TYPE_Q4_K for all tensors |
| wizardlm-13b-v1.1.ggmlv3.q5_0.bin | q5_0 | 5 | 8.95 GB| 11.45 GB | Original quant method, 5-bit. Higher accuracy, higher resource usage and slower inference. |
| wizardlm-13b-v1.1.ggmlv3.q5_1.bin | q5_1 | 5 | 9.76 GB| 12.26 GB | Original quant method, 5-bit. Even higher accuracy, resource usage and slower inference. |
| wizardlm-13b-v1.1.ggmlv3.q5_K_M.bin | q5_K_M | 5 | 9.33 GB| 11.83 GB | New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q5_K |
| wizardlm-13b-v1.1.ggmlv3.q5_K_S.bin | q5_K_S | 5 | 9.07 GB| 11.57 GB | New k-quant method. Uses GGML_TYPE_Q5_K for all tensors |
| wizardlm-13b-v1.1.ggmlv3.q6_K.bin | q6_K | 6 | 10.76 GB| 13.26 GB | New k-quant method. Uses GGML_TYPE_Q8_K - 6-bit quantization - for all tensors |
| wizardlm-13b-v1.1.ggmlv3.q8_0.bin | q8_0 | 8 | 13.83 GB| 16.33 GB | Original quant method, 8-bit. Almost indistinguishable from float16. High resource use and slow. Not recommended for most users. |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
## How to run in `llama.cpp`
I use the following command line; adjust for your tastes and needs:
```
./main -t 10 -ngl 32 -m wizardlm-13b-v1.1.ggmlv3.q4_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "### Instruction: Write a story about llamas\n### Response:"
```
Change `-t 10` to the number of physical CPU cores you have. For example if your system has 8 cores/16 threads, use `-t 8`.
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
## How to run in `text-generation-webui`
Further instructions here: [text-generation-webui/docs/llama.cpp-models.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp-models.md).
<!-- footer start -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Luke from CarbonQuill, Aemon Algiz.
**Patreon special mentions**: RoA, Lone Striker, Gabriel Puliatti, Derek Yates, Randy H, Jonathan Leane, Eugene Pentland, Karl Bernard, Viktor Bowallius, senxiiz, Daniel P. Andersen, Pierre Kircher, Deep Realms, Cory Kujawski, Oscar Rangel, Fen Risland, Ajan Kanaga, LangChain4j, webtim, Nikolai Manek, Trenton Dambrowitz, Raven Klaugh, Kalila, Khalefa Al-Ahmad, Chris McCloskey, Luke @flexchar, Ai Maven, Dave, Asp the Wyvern, Sean Connelly, Imad Khwaja, Space Cruiser, Rainer Wilmers, subjectnull, Alps Aficionado, Willian Hasse, Fred von Graf, Artur Olbinski, Johann-Peter Hartmann, WelcomeToTheClub, Willem Michiel, Michael Levine, Iucharbius , Spiking Neurons AB, K, biorpg, John Villwock, Pyrater, Greatston Gnanesh, Mano Prime, Junyu Yang, Stephen Murray, John Detwiler, Luke Pendergrass, terasurfer , Pieter, zynix , Edmond Seymore, theTransient, Nathan LeClaire, vamX, Kevin Schuppel, Preetika Verma, ya boyyy, Alex , SuperWojo, Ghost , Joseph William Delisle, Matthew Berman, Talal Aujan, chris gileta, Illia Dulskyi.
Thank you to all my generous patrons and donaters!
<!-- footer end -->
# Original model card: WizardLM's WizardLM 13B V1.1
This is the **Full-Weight** of WizardLM-13B V1.1 model.
**Repository**: https://github.com/nlpxucan/WizardLM
**Twitter**: https://twitter.com/WizardLM_AI/status/1677282955490918401
- 🔥🔥🔥 [7/7/2023] We released **WizardLM V1.1** models. The **WizardLM-13B-V1.1** is here ([Demo_13B-V1.1](https://e8a06366ccd1c4d1.gradio.app), [Demo_13B-V1.1_bak-1](https://59da107262a25764.gradio.app), [Demo_13B-V1.1_bak-2](https://dfc5113f66739c80.gradio.app), [Full Model Weight](https://huggingface.co/WizardLM/WizardLM-13B-V1.1)). **WizardLM-7B-V1.1**, **WizardLM-30B-V1.1**, and **WizardLM-65B-V1.1** are coming soon. Please checkout the [Full Model Weights](https://huggingface.co/WizardLM) and [paper](https://arxiv.org/abs/2304.12244).
- 🔥🔥🔥 [7/7/2023] The **WizardLM-13B-V1.1** achieves **6.74** on [MT-Bench Leaderboard](https://chat.lmsys.org/?leaderboard), **86.32%** on [AlpacaEval Leaderboard](https://tatsu-lab.github.io/alpaca_eval/), and **99.3%** on [WizardLM Eval](https://github.com/nlpxucan/WizardLM/blob/main/WizardLM/data/WizardLM_testset.jsonl). (Note: MT-Bench and AlpacaEval are all self-test, will push update and request review. All tests are completed under their official settings.)
|
LoupGarou/WizardCoder-Guanaco-15B-V1.0 | LoupGarou | 2023-07-09T17:24:34Z | 1,506 | 5 | transformers | [
"transformers",
"pytorch",
"gpt_bigcode",
"text-generation",
"en",
"dataset:guanaco",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-07-06T03:17:49Z | ---
language:
- en
datasets:
- guanaco
model_hub_library:
- transformers
license:
- apache-2.0
---
## WizardGuanaco-V1.0 Model Card
The WizardCoder-Guanaco-15B-V1.0 is a language model that combines the strengths of the [WizardCoder](https://huggingface.co/WizardLM/WizardCoder-15B-V1.0) base model and the [openassistant-guanaco](https://huggingface.co/datasets/timdettmers/openassistant-guanaco) dataset for finetuning. The openassistant-guanaco dataset was further trimmed to within 2 standard deviations of token size for input and output pairs and all non-english data has been removed to reduce training size requirements.
# Model Description
This model is built on top of the WizardCoder base model, a large language model known for its impressive capabilities in code related instruction. The WizardCoder base model was further finetuned using QLORA on the openassistant-guanaco dataset to enhance its generative abilities.
However, to ensure more targeted learning and data processing, the dataset was trimmed to within 2 standard deviations of token size for question sets. This process enhances the model's ability to generate more precise and relevant answers, eliminating outliers that could potentially distort the responses. In addition, to focus on English language proficiency, all non-English data has been removed from the Guanaco dataset.
# Intended Use
This model is designed to be used for a wide array of text generation tasks that require understanding and generating English text. The model is expected to perform well in tasks such as answering questions, writing essays, summarizing text, translation, and more. However, given the specific data processing and finetuning done, it might be particularly effective for tasks related to English language question-answering systems.
# Limitations
Despite the powerful capabilities of this model, users should be aware of its limitations. The model's knowledge is up to date only until the time it was trained, and it doesn't know about events in the world after that. It can sometimes produce incorrect or nonsensical responses, as it doesn't understand the text in the same way humans do. It should be used as a tool to assist in generating text and not as a sole source of truth.
# How to use
Here is an example of how to use this model:
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
import time
import torch
class Chatbot:
def __init__(self, model_name):
self.tokenizer = AutoTokenizer.from_pretrained(model_name, padding_side='left')
self.model = AutoModelForCausalLM.from_pretrained(model_name, load_in_4bit=True, torch_dtype=torch.bfloat16)
if self.tokenizer.pad_token_id is None:
self.tokenizer.pad_token_id = self.tokenizer.eos_token_id
def get_response(self, prompt):
inputs = self.tokenizer.encode_plus(prompt, return_tensors="pt", padding='max_length', max_length=100)
if next(self.model.parameters()).is_cuda:
inputs = {name: tensor.to('cuda') for name, tensor in inputs.items()}
start_time = time.time()
tokens = self.model.generate(input_ids=inputs['input_ids'],
attention_mask=inputs['attention_mask'],
pad_token_id=self.tokenizer.pad_token_id,
max_new_tokens=400)
end_time = time.time()
output_tokens = tokens[0][inputs['input_ids'].shape[-1]:]
output = self.tokenizer.decode(output_tokens, skip_special_tokens=True)
time_taken = end_time - start_time
return output, time_taken
def main():
chatbot = Chatbot("LoupGarou/WizardCoder-Guanaco-15B-V1.0")
while True:
user_input = input("Enter your prompt: ")
if user_input.lower() == 'quit':
break
output, time_taken = chatbot.get_response(user_input)
print("\033[33m" + output + "\033[0m")
print("Time taken to process: ", time_taken, "seconds")
print("Exited the program.")
if __name__ == "__main__":
main()
```
# Training Procedure
The base WizardCoder model was finetuned on the openassistant-guanaco dataset using QLORA, which was trimmed to within 2 standard deviations of token size for question sets and randomized. All non-English data was also removed from this finetuning dataset.
## Acknowledgements
This model, WizardCoder-Guanaco-15B-V1.0, is simply building on the efforts of two great teams to evaluate the performance of a combined model with the strengths of the [WizardCoder base model](https://huggingface.co/WizardLM/WizardCoder-15B-V1.0) and the [openassistant-guanaco dataset](https://huggingface.co/datasets/timdettmers/openassistant-guanaco).
A sincere appreciation goes out to the developers and the community involved in the creation and refinement of these models. Their commitment to providing open source tools and datasets have been instrumental in making this project a reality.
Moreover, a special note of thanks to the [Hugging Face](https://huggingface.co/) team, whose transformative library has not only streamlined the process of model creation and adaptation, but also democratized the access to state-of-the-art machine learning technologies. Their impact on the development of this project cannot be overstated.
|
guaguale/model_kthv_v1 | guaguale | 2023-07-09T17:11:11Z | 1 | 0 | diffusers | [
"diffusers",
"tensorboard",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"dreambooth",
"base_model:runwayml/stable-diffusion-v1-5",
"base_model:finetune:runwayml/stable-diffusion-v1-5",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] | text-to-image | 2023-07-09T12:03:40Z |
---
license: creativeml-openrail-m
base_model: runwayml/stable-diffusion-v1-5
instance_prompt: a male idol sks with blonde hair, wearing a black jacket and fringes on the sides of the jacket
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- dreambooth
inference: true
---
# DreamBooth - guaguale/model_kthv_v1
This is a dreambooth model derived from runwayml/stable-diffusion-v1-5. The weights were trained on a male idol sks with blonde hair, wearing a black jacket and fringes on the sides of the jacket using [DreamBooth](https://dreambooth.github.io/).
You can find some example images in the following.
DreamBooth for the text encoder was enabled: False.
|
digiplay/LuckyStrikeMix0.2Realistic | digiplay | 2023-07-09T17:07:13Z | 311 | 1 | diffusers | [
"diffusers",
"safetensors",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"license:other",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] | text-to-image | 2023-07-08T13:03:32Z | ---
license: other
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
inference: true
---
Model info :
https://civitai.com/models/13034/lucky-strike-mix
***Note: please use "realistic" keywords to make some realistic results.***
Sample image I made thru huggingface's API:
```
realistic ,MCU,(masterpiece, best quality, ultra high res:1.3), 1girl, (abstract art:1.3), half demon,
```
|
RogerB/afriberta_small-finetuned-kintweetsD | RogerB | 2023-07-09T17:05:25Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"fill-mask",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-07-09T16:55:42Z | ---
tags:
- generated_from_trainer
model-index:
- name: afriberta_small-finetuned-kintweetsD
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. -->
# afriberta_small-finetuned-kintweetsD
This model is a fine-tuned version of [castorini/afriberta_small](https://huggingface.co/castorini/afriberta_small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.2332
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.6431 | 1.0 | 900 | 3.3421 |
| 3.4111 | 2.0 | 1800 | 3.2661 |
| 3.3391 | 3.0 | 2700 | 3.2382 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
ericNguyen0132/roberta-large-Dep-second | ericNguyen0132 | 2023-07-09T16:54:58Z | 3 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"roberta",
"text-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | text-classification | 2023-07-07T16:25:28Z | ---
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: roberta-large-Dep-second
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-large-Dep-second
This model is a fine-tuned version of [rafalposwiata/deproberta-large-depression](https://huggingface.co/rafalposwiata/deproberta-large-depression) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1600
- Accuracy: 0.8517
- F1: 0.9113
## 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-06
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| No log | 1.0 | 469 | 0.3551 | 0.86 | 0.9188 |
| 0.3676 | 2.0 | 938 | 0.4666 | 0.8617 | 0.9198 |
| 0.3042 | 3.0 | 1407 | 0.5818 | 0.86 | 0.9170 |
| 0.2651 | 4.0 | 1876 | 0.8291 | 0.865 | 0.9200 |
| 0.174 | 5.0 | 2345 | 0.8843 | 0.8567 | 0.9155 |
| 0.1363 | 6.0 | 2814 | 1.1669 | 0.8317 | 0.8968 |
| 0.075 | 7.0 | 3283 | 1.2803 | 0.8283 | 0.8952 |
| 0.0401 | 8.0 | 3752 | 1.0247 | 0.8617 | 0.9184 |
| 0.0301 | 9.0 | 4221 | 1.2848 | 0.83 | 0.8961 |
| 0.0281 | 10.0 | 4690 | 1.1600 | 0.8517 | 0.9113 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
turhancan97/yolov8-segment-trash-detection | turhancan97 | 2023-07-09T16:53:38Z | 0 | 1 | null | [
"computer-vision",
"yolov8",
"segmentation",
"en",
"dataset:garythung/trashnet",
"dataset:Zesky665/TACO",
"dataset:detection-datasets/coco",
"license:mit",
"region:us"
] | null | 2023-07-09T16:42:33Z | ---
license: mit
language:
- en
datasets:
- garythung/trashnet
- Zesky665/TACO
- detection-datasets/coco
tags:
- computer-vision
- yolov8
- segmentation
---
### Model Description
[Ultralytics:](https://github.com/ultralytics/ultralytics/) YOLOv8 in PyTorch > ONNX > CoreML > TFLite]
### Installation
```
pip install ultralytics
```
### Yolov8 Inference
```python
from ultralytics import YOLO
model = YOLO('turhancan97/yolov8-segment-trash-detection')
prediction = model.predict(image, imgsz=image_size, show=False, save=False)
``` |
NasimB/gpt2-concat-guten-rarity-iroder-est-rarity-all-5k-2p5k | NasimB | 2023-07-09T16:50:17Z | 5 | 0 | transformers | [
"transformers",
"pytorch",
"gpt2",
"text-generation",
"generated_from_trainer",
"dataset:generator",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-07-09T14:22:52Z | ---
license: mit
tags:
- generated_from_trainer
datasets:
- generator
model-index:
- name: gpt2-concat-guten-rarity-iroder-est-rarity-all-5k-2p5k
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. -->
# gpt2-concat-guten-rarity-iroder-est-rarity-all-5k-2p5k
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the generator dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1825
## 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: 0.0005
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 6
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:-----:|:---------------:|
| 6.7149 | 0.3 | 500 | 5.6379 |
| 5.3626 | 0.59 | 1000 | 5.2053 |
| 5.0103 | 0.89 | 1500 | 4.9561 |
| 4.7374 | 1.18 | 2000 | 4.8042 |
| 4.5794 | 1.48 | 2500 | 4.6786 |
| 4.4756 | 1.77 | 3000 | 4.5728 |
| 4.3345 | 2.07 | 3500 | 4.4966 |
| 4.1538 | 2.36 | 4000 | 4.4468 |
| 4.1275 | 2.66 | 4500 | 4.3875 |
| 4.0815 | 2.95 | 5000 | 4.3316 |
| 3.8534 | 3.25 | 5500 | 4.3293 |
| 3.8151 | 3.54 | 6000 | 4.2939 |
| 3.8069 | 3.84 | 6500 | 4.2572 |
| 3.6706 | 4.13 | 7000 | 4.2627 |
| 3.5322 | 4.43 | 7500 | 4.2534 |
| 3.5274 | 4.73 | 8000 | 4.2397 |
| 3.4978 | 5.02 | 8500 | 4.2343 |
| 3.3402 | 5.32 | 9000 | 4.2407 |
| 3.3342 | 5.61 | 9500 | 4.2388 |
| 3.3308 | 5.91 | 10000 | 4.2382 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
|
Hedayat-Abrishami/ppo-SnowballTarget | Hedayat-Abrishami | 2023-07-09T16:37:35Z | 3 | 0 | ml-agents | [
"ml-agents",
"tensorboard",
"onnx",
"SnowballTarget",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-SnowballTarget",
"region:us"
] | reinforcement-learning | 2023-07-09T16:37:33Z | ---
library_name: ml-agents
tags:
- SnowballTarget
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-SnowballTarget
---
# **ppo** Agent playing **SnowballTarget**
This is a trained model of a **ppo** agent playing **SnowballTarget**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: Hedayat-Abrishami/ppo-SnowballTarget
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
hoanghoavienvo/bert-base-uncased-detect-depression-stage-one | hoanghoavienvo | 2023-07-09T16:29:45Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"bert",
"text-classification",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | text-classification | 2023-07-09T14:53:19Z | ---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: bert-base-uncased-detect-depression-stage-one
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. -->
# bert-base-uncased-detect-depression-stage-one
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4151
- Accuracy: 0.674
- F1: 0.7530
## 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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.6185 | 1.0 | 751 | 0.5223 | 0.762 | 0.8260 |
| 0.5394 | 2.0 | 1502 | 0.5958 | 0.692 | 0.7488 |
| 0.4387 | 3.0 | 2253 | 0.6742 | 0.706 | 0.7885 |
| 0.3157 | 4.0 | 3004 | 1.1284 | 0.696 | 0.7768 |
| 0.1794 | 5.0 | 3755 | 1.4151 | 0.674 | 0.7530 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
Daniil-plotnikov/russian-vision-v3 | Daniil-plotnikov | 2023-07-09T16:21:11Z | 33 | 0 | diffusers | [
"diffusers",
"text-to-image",
"stable-diffusion",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] | text-to-image | 2023-07-09T16:09:29Z | ---
license: creativeml-openrail-m
tags:
- text-to-image
- stable-diffusion
---
### Russian-Vision-V3 Dreambooth model trained by Daniil-plotnikov with [TheLastBen's fast-DreamBooth](https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast-DreamBooth.ipynb) notebook
Test the concept via A1111 Colab [fast-Colab-A1111](https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast_stable_diffusion_AUTOMATIC1111.ipynb)
Sample pictures of this concept:
|
EleutherAI/pythia-70m-deduped | EleutherAI | 2023-07-09T16:07:33Z | 122,668 | 25 | transformers | [
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"en",
"dataset:EleutherAI/the_pile_deduplicated",
"arxiv:2304.01373",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-02-13T16:01:41Z | ---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
license: apache-2.0
datasets:
- EleutherAI/the_pile_deduplicated
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf).
It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. We also provide 154 intermediate
checkpoints per model, hosted on Hugging Face as branches.
The Pythia model suite was designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
<details>
<summary style="font-weight:600">Details on previous early release and naming convention.</summary>
Previously, we released an early version of the Pythia suite to the public.
However, we decided to retrain the model suite to address a few hyperparameter
discrepancies. This model card <a href="#changelog">lists the changes</a>;
see appendix B in the Pythia paper for further discussion. We found no
difference in benchmark performance between the two Pythia versions.
The old models are
[still available](https://huggingface.co/models?other=pythia_v0), but we
suggest the retrained suite if you are just starting to use Pythia.<br>
**This is the current release.**
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
</details>
<br>
# Pythia-70M-deduped
## Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
[See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation
details.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:[email protected]).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
## Uses and Limitations
### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. We also provide
154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints
`step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to
`step143000`. These checkpoints are hosted on Hugging Face as branches. Note
that branch `143000` corresponds exactly to the model checkpoint on the `main`
branch of each model.
You may also further fine-tune and adapt Pythia-70M-deduped for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-70M-deduped as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions. For example,
the model may generate harmful or offensive text. Please evaluate the risks
associated with your particular use case.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-70M-deduped has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-70M-deduped will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “follow” human instructions.
### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token used by the model need not produce the
most “accurate” text. Never rely on Pythia-70M-deduped to produce factually accurate
output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-70M-deduped may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-70M-deduped.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
## Training
### Training data
Pythia-70M-deduped was trained on the Pile **after the dataset has been globally
deduplicated**.<br>
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).
### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training,
from `step1000` to `step143000` (which is the same as `main`). In addition, we
also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for 143000 steps at a batch size
of 2M (2,097,152 tokens).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
## Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Easy Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/>
</details>
## Changelog
This section compares differences between previously released
[Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current
models. See Appendix B of the Pythia paper for further discussion of these
changes and the motivation behind them. We found that retraining Pythia had no
impact on benchmark performance.
- All model sizes are now trained with uniform batch size of 2M tokens.
Previously, the models of size 160M, 410M, and 1.4B parameters were trained
with batch sizes of 4M tokens.
- We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64,
128,256,512} in addition to every 1000 training steps.
- Flash Attention was used in the new retrained suite.
- We remedied a minor inconsistency that existed in the original suite: all
models of size 2.8B parameters or smaller had a learning rate (LR) schedule
which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and
12B models all used an LR schedule which decayed to a minimum LR of 0. In
the redone training runs, we rectified this inconsistency: all models now were
trained with LR decaying to a minimum of 0.1× their maximum LR.
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure> |
EleutherAI/pythia-160m-deduped | EleutherAI | 2023-07-09T16:04:57Z | 43,492 | 3 | transformers | [
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"en",
"dataset:EleutherAI/the_pile_deduplicated",
"arxiv:2304.01373",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-02-08T21:50:19Z | ---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
license: apache-2.0
datasets:
- EleutherAI/the_pile_deduplicated
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf).
It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. We also provide 154 intermediate
checkpoints per model, hosted on Hugging Face as branches.
The Pythia model suite was designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
<details>
<summary style="font-weight:600">Details on previous early release and naming convention.</summary>
Previously, we released an early version of the Pythia suite to the public.
However, we decided to retrain the model suite to address a few hyperparameter
discrepancies. This model card <a href="#changelog">lists the changes</a>;
see appendix B in the Pythia paper for further discussion. We found no
difference in benchmark performance between the two Pythia versions.
The old models are
[still available](https://huggingface.co/models?other=pythia_v0), but we
suggest the retrained suite if you are just starting to use Pythia.<br>
**This is the current release.**
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
</details>
<br>
# Pythia-160M-deduped
## Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
[See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation
details.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:[email protected]).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
## Uses and Limitations
### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. We also provide
154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints
`step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to
`step143000`. These checkpoints are hosted on Hugging Face as branches. Note
that branch `143000` corresponds exactly to the model checkpoint on the `main`
branch of each model.
You may also further fine-tune and adapt Pythia-160M-deduped for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-160M-deduped as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions. For example,
the model may generate harmful or offensive text. Please evaluate the risks
associated with your particular use case.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-160M-deduped has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-160M-deduped will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “follow” human instructions.
### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token used by the model need not produce the
most “accurate” text. Never rely on Pythia-160M-deduped to produce factually accurate
output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-160M-deduped may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-160M-deduped.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
## Training
### Training data
Pythia-160M-deduped was trained on the Pile **after the dataset has been globally
deduplicated**.<br>
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).
### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training,
from `step1000` to `step143000` (which is the same as `main`). In addition, we
also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for 143000 steps at a batch size
of 2M (2,097,152 tokens).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
## Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Easy Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/>
</details>
## Changelog
This section compares differences between previously released
[Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current
models. See Appendix B of the Pythia paper for further discussion of these
changes and the motivation behind them. We found that retraining Pythia had no
impact on benchmark performance.
- All model sizes are now trained with uniform batch size of 2M tokens.
Previously, the models of size 160M, 410M, and 1.4B parameters were trained
with batch sizes of 4M tokens.
- We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64,
128,256,512} in addition to every 1000 training steps.
- Flash Attention was used in the new retrained suite.
- We remedied a minor inconsistency that existed in the original suite: all
models of size 2.8B parameters or smaller had a learning rate (LR) schedule
which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and
12B models all used an LR schedule which decayed to a minimum LR of 0. In
the redone training runs, we rectified this inconsistency: all models now were
trained with LR decaying to a minimum of 0.1× their maximum LR.
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure> |
EleutherAI/pythia-1.4b | EleutherAI | 2023-07-09T16:01:57Z | 25,256 | 22 | transformers | [
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"en",
"dataset:EleutherAI/the_pile",
"arxiv:2304.01373",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-02-09T14:08:20Z | ---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
license: apache-2.0
datasets:
- EleutherAI/the_pile
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf).
It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. We also provide 154 intermediate
checkpoints per model, hosted on Hugging Face as branches.
The Pythia model suite was deliberately designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
<details>
<summary style="font-weight:600">Details on previous early release and naming convention.</summary>
Previously, we released an early version of the Pythia suite to the public.
However, we decided to retrain the model suite to address a few hyperparameter
discrepancies. This model card <a href="#changelog">lists the changes</a>;
see appendix B in the Pythia paper for further discussion. We found no
difference in benchmark performance between the two Pythia versions.
The old models are
[still available](https://huggingface.co/models?other=pythia_v0), but we
suggest the retrained suite if you are just starting to use Pythia.<br>
**This is the current release.**
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
</details>
<br>
# Pythia-1.4B
## Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
[See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation
details.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:[email protected]).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
## Uses and Limitations
### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. We also provide
154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints
`step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to
`step143000`. These checkpoints are hosted on Hugging Face as branches. Note
that branch `143000` corresponds exactly to the model checkpoint on the `main`
branch of each model.
You may also further fine-tune and adapt Pythia-1.4B for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-1.4B as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions. For example,
the model may generate harmful or offensive text. Please evaluate the risks
associated with your particular use case.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-1.4B has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-1.4B will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “follow” human instructions.
### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token used by the model need not produce the
most “accurate” text. Never rely on Pythia-1.4B to produce factually accurate
output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-1.4B may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-1.4B.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
## Training
### Training data
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).<br>
The Pile was **not** deduplicated before being used to train Pythia-1.4B.
### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training,
from `step1000` to `step143000` (which is the same as `main`). In addition, we
also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for 143000 steps at a batch size
of 2M (2,097,152 tokens).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
## Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Easy Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/>
</details>
## Changelog
This section compares differences between previously released
[Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current
models. See Appendix B of the Pythia paper for further discussion of these
changes and the motivation behind them. We found that retraining Pythia had no
impact on benchmark performance.
- All model sizes are now trained with uniform batch size of 2M tokens.
Previously, the models of size 160M, 410M, and 1.4B parameters were trained
with batch sizes of 4M tokens.
- We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64,
128,256,512} in addition to every 1000 training steps.
- Flash Attention was used in the new retrained suite.
- We remedied a minor inconsistency that existed in the original suite: all
models of size 2.8B parameters or smaller had a learning rate (LR) schedule
which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and
12B models all used an LR schedule which decayed to a minimum LR of 0. In
the redone training runs, we rectified this inconsistency: all models now were
trained with LR decaying to a minimum of 0.1× their maximum LR.
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure> |
EleutherAI/pythia-410m | EleutherAI | 2023-07-09T16:01:42Z | 68,125 | 22 | transformers | [
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"causal-lm",
"pythia",
"en",
"dataset:EleutherAI/pile",
"arxiv:2304.01373",
"arxiv:2101.00027",
"arxiv:2201.07311",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-02-13T18:45:00Z | ---
language:
- en
tags:
- pytorch
- causal-lm
- pythia
license: apache-2.0
datasets:
- EleutherAI/pile
---
The *Pythia Scaling Suite* is a collection of models developed to facilitate
interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf).
It contains two sets of eight models of sizes
70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two
models: one trained on the Pile, and one trained on the Pile after the dataset
has been globally deduplicated. All 8 model sizes are trained on the exact
same data, in the exact same order. We also provide 154 intermediate
checkpoints per model, hosted on Hugging Face as branches.
The Pythia model suite was deliberately designed to promote scientific
research on large language models, especially interpretability research.
Despite not centering downstream performance as a design goal, we find the
models <a href="#evaluations">match or exceed</a> the performance of
similar and same-sized models, such as those in the OPT and GPT-Neo suites.
<details>
<summary style="font-weight:600">Details on previous early release and naming convention.</summary>
Previously, we released an early version of the Pythia suite to the public.
However, we decided to retrain the model suite to address a few hyperparameter
discrepancies. This model card <a href="#changelog">lists the changes</a>;
see appendix B in the Pythia paper for further discussion. We found no
difference in benchmark performance between the two Pythia versions.
The old models are
[still available](https://huggingface.co/models?other=pythia_v0), but we
suggest the retrained suite if you are just starting to use Pythia.<br>
**This is the current release.**
Please note that all models in the *Pythia* suite were renamed in January
2023. For clarity, a <a href="#naming-convention-and-parameter-count">table
comparing the old and new names</a> is provided in this model card, together
with exact parameter counts.
</details>
<br>
# Pythia-410M
## Model Details
- Developed by: [EleutherAI](http://eleuther.ai)
- Model type: Transformer-based Language Model
- Language: English
- Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia)
for training procedure, config files, and details on how to use.
[See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation
details.
- Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
- License: Apache 2.0
- Contact: to ask questions about this model, join the [EleutherAI
Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`.
Please read the existing *Pythia* documentation before asking about it in the
EleutherAI Discord. For general correspondence: [contact@eleuther.
ai](mailto:[email protected]).
<figure>
| Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models |
| -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: |
| 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10<sup>-3</sup> | — |
| 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10<sup>-4</sup> | GPT-Neo 125M, OPT-125M |
| 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10<sup>-4</sup> | OPT-350M |
| 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10<sup>-4</sup> | — |
| 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10<sup>-4</sup> | GPT-Neo 1.3B, OPT-1.3B |
| 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10<sup>-4</sup> | GPT-Neo 2.7B, OPT-2.7B |
| 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10<sup>-4</sup> | OPT-6.7B |
| 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10<sup>-4</sup> | — |
<figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and
non-deduped models of a given size have the same hyperparameters. “Equivalent”
models have <b>exactly</b> the same architecture, and the same number of
non-embedding parameters.</figcaption>
</figure>
## Uses and Limitations
### Intended Use
The primary intended use of Pythia is research on the behavior, functionality,
and limitations of large language models. This suite is intended to provide
a controlled setting for performing scientific experiments. We also provide
154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints
`step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to
`step143000`. These checkpoints are hosted on Hugging Face as branches. Note
that branch `143000` corresponds exactly to the model checkpoint on the `main`
branch of each model.
You may also further fine-tune and adapt Pythia-410M for deployment,
as long as your use is in accordance with the Apache 2.0 license. Pythia
models work with the Hugging Face [Transformers
Library](https://huggingface.co/docs/transformers/index). If you decide to use
pre-trained Pythia-410M as a basis for your fine-tuned model, please
conduct your own risk and bias assessment.
### Out-of-scope use
The Pythia Suite is **not** intended for deployment. It is not a in itself
a product and cannot be used for human-facing interactions. For example,
the model may generate harmful or offensive text. Please evaluate the risks
associated with your particular use case.
Pythia models are English-language only, and are not suitable for translation
or generating text in other languages.
Pythia-410M has not been fine-tuned for downstream contexts in which
language models are commonly deployed, such as writing genre prose,
or commercial chatbots. This means Pythia-410M will **not**
respond to a given prompt the way a product like ChatGPT does. This is because,
unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement
Learning from Human Feedback (RLHF) to better “follow” human instructions.
### Limitations and biases
The core functionality of a large language model is to take a string of text
and predict the next token. The token used by the model need not produce the
most “accurate” text. Never rely on Pythia-410M to produce factually accurate
output.
This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset
known to contain profanity and texts that are lewd or otherwise offensive.
See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a
discussion of documented biases with regards to gender, religion, and race.
Pythia-410M may produce socially unacceptable or undesirable text, *even if*
the prompt itself does not include anything explicitly offensive.
If you plan on using text generated through, for example, the Hosted Inference
API, we recommend having a human curate the outputs of this language model
before presenting it to other people. Please inform your audience that the
text was generated by Pythia-410M.
### Quickstart
Pythia models can be loaded and used via the following code, demonstrated here
for the third `pythia-70m-deduped` checkpoint:
```python
from transformers import GPTNeoXForCausalLM, AutoTokenizer
model = GPTNeoXForCausalLM.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
tokenizer = AutoTokenizer.from_pretrained(
"EleutherAI/pythia-70m-deduped",
revision="step3000",
cache_dir="./pythia-70m-deduped/step3000",
)
inputs = tokenizer("Hello, I am", return_tensors="pt")
tokens = model.generate(**inputs)
tokenizer.decode(tokens[0])
```
Revision/branch `step143000` corresponds exactly to the model checkpoint on
the `main` branch of each model.<br>
For more information on how to use all Pythia models, see [documentation on
GitHub](https://github.com/EleutherAI/pythia).
## Training
### Training data
[The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in
English. It was created by EleutherAI specifically for training large language
models. It contains texts from 22 diverse sources, roughly broken down into
five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl),
prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and
miscellaneous (e.g. GitHub, Enron Emails). See [the Pile
paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources,
methodology, and a discussion of ethical implications. Consult [the
datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation
about the Pile and its component datasets. The Pile can be downloaded from
the [official website](https://pile.eleuther.ai/), or from a [community
mirror](https://the-eye.eu/public/AI/pile/).<br>
The Pile was **not** deduplicated before being used to train Pythia-410M.
### Training procedure
All models were trained on the exact same data, in the exact same order. Each
model saw 299,892,736,000 tokens during training, and 143 checkpoints for each
model are saved every 2,097,152,000 tokens, spaced evenly throughout training,
from `step1000` to `step143000` (which is the same as `main`). In addition, we
also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`.
This corresponds to training for just under 1 epoch on the Pile for
non-deduplicated models, and about 1.5 epochs on the deduplicated Pile.
All *Pythia* models trained for 143000 steps at a batch size
of 2M (2,097,152 tokens).<br>
See [GitHub](https://github.com/EleutherAI/pythia) for more details on training
procedure, including [how to reproduce
it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br>
Pythia uses the same tokenizer as [GPT-NeoX-
20B](https://huggingface.co/EleutherAI/gpt-neox-20b).
## Evaluations
All 16 *Pythia* models were evaluated using the [LM Evaluation
Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access
the results by model and step at `results/json/*` in the [GitHub
repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br>
Expand the sections below to see plots of evaluation results for all
Pythia and Pythia-deduped models compared with OPT and BLOOM.
<details>
<summary>LAMBADA – OpenAI</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/>
</details>
<details>
<summary>Physical Interaction: Question Answering (PIQA)</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/>
</details>
<details>
<summary>WinoGrande</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/>
</details>
<details>
<summary>AI2 Reasoning Challenge—Easy Set</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/>
</details>
<details>
<summary>SciQ</summary>
<img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/>
</details>
## Changelog
This section compares differences between previously released
[Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current
models. See Appendix B of the Pythia paper for further discussion of these
changes and the motivation behind them. We found that retraining Pythia had no
impact on benchmark performance.
- All model sizes are now trained with uniform batch size of 2M tokens.
Previously, the models of size 160M, 410M, and 1.4B parameters were trained
with batch sizes of 4M tokens.
- We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64,
128,256,512} in addition to every 1000 training steps.
- Flash Attention was used in the new retrained suite.
- We remedied a minor inconsistency that existed in the original suite: all
models of size 2.8B parameters or smaller had a learning rate (LR) schedule
which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and
12B models all used an LR schedule which decayed to a minimum LR of 0. In
the redone training runs, we rectified this inconsistency: all models now were
trained with LR decaying to a minimum of 0.1× their maximum LR.
### Naming convention and parameter count
*Pythia* models were renamed in January 2023. It is possible that the old
naming convention still persists in some documentation by accident. The
current naming convention (70M, 160M, etc.) is based on total parameter count.
<figure style="width:32em">
| current Pythia suffix | old suffix | total params | non-embedding params |
| --------------------: | ---------: | -------------: | -------------------: |
| 70M | 19M | 70,426,624 | 18,915,328 |
| 160M | 125M | 162,322,944 | 85,056,000 |
| 410M | 350M | 405,334,016 | 302,311,424 |
| 1B | 800M | 1,011,781,632 | 805,736,448 |
| 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 |
| 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 |
| 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 |
| 12B | 13B | 11,846,072,320 | 11,327,027,200 |
</figure> |
cagarraz/poca-SoccerTwos | cagarraz | 2023-07-09T16:01:25Z | 0 | 0 | ml-agents | [
"ml-agents",
"tensorboard",
"SoccerTwos",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-SoccerTwos",
"region:us"
] | reinforcement-learning | 2023-07-09T16:01:19Z | ---
library_name: ml-agents
tags:
- SoccerTwos
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-SoccerTwos
---
# **poca** Agent playing **SoccerTwos**
This is a trained model of a **poca** agent playing **SoccerTwos**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: cagarraz/poca-SoccerTwos
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
EleutherAI/pythia-intervention-410m-deduped | EleutherAI | 2023-07-09T16:00:37Z | 28 | 1 | transformers | [
"transformers",
"pytorch",
"safetensors",
"gpt_neox",
"text-generation",
"dataset:EleutherAI/pile",
"arxiv:2304.01373",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-03-03T14:52:01Z | ---
license: apache-2.0
datasets:
- EleutherAI/pile
---
This model is part of an intervention study done in the paper [Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling](https://arxiv.org/abs/2304.01373) where we replaced all masculine pronouns with femanine ones and retrained the model for the last 21 billion tokens. The regular model can be found [here](https://huggingface.co/EleutherAI/pythia-410m-deduped).
**We do not recommend using this model for any purpose other than to study the influence of gender pronouns on language model behavior.** |
agercas/whisper-tiny-us | agercas | 2023-07-09T15:58:24Z | 77 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"whisper",
"automatic-speech-recognition",
"generated_from_trainer",
"dataset:PolyAI/minds14",
"license:apache-2.0",
"model-index",
"endpoints_compatible",
"region:us"
] | automatic-speech-recognition | 2023-07-09T14:56:53Z | ---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- PolyAI/minds14
metrics:
- wer
model-index:
- name: whisper-tiny-us
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: PolyAI/minds14
type: PolyAI/minds14
config: en-US
split: train
args: en-US
metrics:
- name: Wer
type: wer
value: 0.33116883116883117
---
<!-- 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. -->
# whisper-tiny-us
This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the PolyAI/minds14 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7183
- Wer Ortho: 0.3381
- Wer: 0.3312
## 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: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 50
- training_steps: 500
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|
| 0.0012 | 17.86 | 500 | 0.7183 | 0.3381 | 0.3312 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
RogerB/afro-xlmr-mini-finetuned-kintweetsD | RogerB | 2023-07-09T15:57:33Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"fill-mask",
"generated_from_trainer",
"license:afl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-07-09T15:40:14Z | ---
license: afl-3.0
tags:
- generated_from_trainer
model-index:
- name: afro-xlmr-mini-finetuned-kintweetsD
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. -->
# afro-xlmr-mini-finetuned-kintweetsD
This model is a fine-tuned version of [Davlan/afro-xlmr-mini](https://huggingface.co/Davlan/afro-xlmr-mini) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.7420
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.1414 | 1.0 | 900 | 2.8278 |
| 2.9859 | 2.0 | 1800 | 2.7708 |
| 2.9245 | 3.0 | 2700 | 2.7253 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
carybr/ppo-Huggy | carybr | 2023-07-09T15:37:56Z | 9 | 0 | ml-agents | [
"ml-agents",
"tensorboard",
"onnx",
"Huggy",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-Huggy",
"region:us"
] | reinforcement-learning | 2023-07-09T15:37:52Z | ---
library_name: ml-agents
tags:
- Huggy
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-Huggy
---
# **ppo** Agent playing **Huggy**
This is a trained model of a **ppo** agent playing **Huggy**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: carybr/ppo-Huggy
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
afterthougt/kullm-polyglot-12.8b-v2_700steps | afterthougt | 2023-07-09T15:31:45Z | 5 | 0 | peft | [
"peft",
"gpt_neox",
"region:us"
] | null | 2023-07-06T05:04:33Z | ---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: bfloat16
### Framework versions
- PEFT 0.4.0.dev0
|
lordtt13/emo-mobilebert | lordtt13 | 2023-07-09T15:28:20Z | 3,195 | 3 | transformers | [
"transformers",
"pytorch",
"tf",
"safetensors",
"mobilebert",
"text-classification",
"en",
"dataset:emo",
"arxiv:2004.02984",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | text-classification | 2022-03-02T23:29:05Z | ---
language: en
datasets:
- emo
---
## Emo-MobileBERT: a thin version of BERT LARGE, trained on the EmoContext Dataset from scratch
### Details of MobileBERT
The **MobileBERT** model was presented in [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by *Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, Denny Zhou* and here is the abstract:
Natural Language Processing (NLP) has recently achieved great success by using huge pre-trained models with hundreds of millions of parameters. However, these models suffer from heavy model sizes and high latency such that they cannot be deployed to resource-limited mobile devices. In this paper, we propose MobileBERT for compressing and accelerating the popular BERT model. Like the original BERT, MobileBERT is task-agnostic, that is, it can be generically applied to various downstream NLP tasks via simple fine-tuning. Basically, MobileBERT is a thin version of BERT_LARGE, while equipped with bottleneck structures and a carefully designed balance between self-attentions and feed-forward networks. To train MobileBERT, we first train a specially designed teacher model, an inverted-bottleneck incorporated BERT_LARGE model. Then, we conduct knowledge transfer from this teacher to MobileBERT. Empirical studies show that MobileBERT is 4.3x smaller and 5.5x faster than BERT_BASE while achieving competitive results on well-known benchmarks. On the natural language inference tasks of GLUE, MobileBERT achieves a GLUEscore o 77.7 (0.6 lower than BERT_BASE), and 62 ms latency on a Pixel 4 phone. On the SQuAD v1.1/v2.0 question answering task, MobileBERT achieves a dev F1 score of 90.0/79.2 (1.5/2.1 higher than BERT_BASE).
### Details of the downstream task (Emotion Recognition) - Dataset 📚
SemEval-2019 Task 3: EmoContext Contextual Emotion Detection in Text
In this dataset, given a textual dialogue i.e. an utterance along with two previous turns of context, the goal was to infer the underlying emotion of the utterance by choosing from four emotion classes:
- sad 😢
- happy 😃
- angry 😡
- others
### Model training
The training script is present [here](https://github.com/lordtt13/transformers-experiments/blob/master/Custom%20Tasks/emo-mobilebert.ipynb).
### Pipelining the Model
```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline
tokenizer = AutoTokenizer.from_pretrained("lordtt13/emo-mobilebert")
model = AutoModelForSequenceClassification.from_pretrained("lordtt13/emo-mobilebert")
nlp_sentence_classif = transformers.pipeline('sentiment-analysis', model = model, tokenizer = tokenizer)
nlp_sentence_classif("I've never had such a bad day in my life")
# Output: [{'label': 'sad', 'score': 0.93153977394104}]
```
> Created by [Tanmay Thakur](https://github.com/lordtt13) | [LinkedIn](https://www.linkedin.com/in/tanmay-thakur-6bb5a9154/)
|
Sekiraw/doom | Sekiraw | 2023-07-09T15:12:54Z | 0 | 0 | stable-baselines3 | [
"stable-baselines3",
"VizdoomBasic-v0",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] | reinforcement-learning | 2023-07-09T15:07:55Z | ---
library_name: stable-baselines3
tags:
- VizdoomBasic-v0
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: ppo
results:
- metrics:
- type: mean_reward
value: -209.90 +/- 196.26
name: mean_reward
task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: VizdoomBasic-v0
type: VizdoomBasic-v0
---
# **ppo** Agent playing **VizdoomBasic-v0**
This is a trained model of a **ppo** agent playing **VizdoomBasic-v0** using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
|
Gorttham/flan-t5-small-chat | Gorttham | 2023-07-09T15:07:44Z | 111 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"t5",
"text2text-generation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text2text-generation | 2023-07-02T10:11:21Z | ---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: content
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. -->
# content
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.5985
- Rouge1: 32.7607
- Rouge2: 19.5507
- Rougel: 32.7312
- Rougelsum: 32.7306
- Gen Len: 16.4212
## 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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 3.043 | 1.0 | 1348 | 2.7636 | 34.4245 | 20.9962 | 34.386 | 34.3876 | 15.1150 |
| 2.8078 | 2.0 | 2696 | 2.6540 | 32.5342 | 19.3983 | 32.4966 | 32.4947 | 16.7662 |
| 2.7166 | 3.0 | 4044 | 2.6103 | 32.4564 | 19.3597 | 32.4255 | 32.4355 | 16.6037 |
| 2.6876 | 4.0 | 5392 | 2.5985 | 32.7607 | 19.5507 | 32.7312 | 32.7306 | 16.4212 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
chunwoolee0/my_awesome_eli5_clm-model | chunwoolee0 | 2023-07-09T15:06:15Z | 141 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"gpt2",
"text-generation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-07-09T11:57:24Z | ---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: my_awesome_eli5_clm-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. -->
# my_awesome_eli5_clm-model
This model is a fine-tuned version of [distilgpt2](https://huggingface.co/distilgpt2) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.7493
## 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: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.7059 | 1.0 | 1108 | 3.7527 |
| 3.6588 | 2.0 | 2216 | 3.7516 |
| 3.6291 | 3.0 | 3324 | 3.7493 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
MostafaHamwi/TextSimplification | MostafaHamwi | 2023-07-09T15:05:20Z | 63 | 0 | transformers | [
"transformers",
"tf",
"t5",
"text2text-generation",
"generated_from_keras_callback",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | text2text-generation | 2023-06-29T23:53:00Z | ---
license: apache-2.0
tags:
- generated_from_keras_callback
model-index:
- name: TextSimplification
results: []
---
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# TextSimplification
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on D-Wikiepdia 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:
- optimizer: None
- training_precision: float32
### Training results
### Framework versions
- Transformers 4.30.2
- TensorFlow 2.12.0
- Tokenizers 0.13.3
|
whiteDandelion/xlm-roberta-base-finetuned-panx-de | whiteDandelion | 2023-07-09T15:05:10Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"token-classification",
"generated_from_trainer",
"dataset:xtreme",
"license:mit",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | token-classification | 2023-07-09T14:54:31Z | ---
license: mit
tags:
- generated_from_trainer
datasets:
- xtreme
metrics:
- f1
model-index:
- name: xlm-roberta-base-finetuned-panx-de
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: xtreme
type: xtreme
config: PAN-X.de
split: validation
args: PAN-X.de
metrics:
- name: F1
type: f1
value: 0.8653353814644136
---
<!-- 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. -->
# xlm-roberta-base-finetuned-panx-de
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the xtreme dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1339
- F1: 0.8653
## 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: 24
- eval_batch_size: 24
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.2583 | 1.0 | 525 | 0.1596 | 0.8231 |
| 0.1262 | 2.0 | 1050 | 0.1395 | 0.8468 |
| 0.0824 | 3.0 | 1575 | 0.1339 | 0.8653 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
agercas/whisper-small-dv | agercas | 2023-07-09T14:43:02Z | 79 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"whisper",
"automatic-speech-recognition",
"generated_from_trainer",
"dv",
"dataset:mozilla-foundation/common_voice_13_0",
"license:apache-2.0",
"model-index",
"endpoints_compatible",
"region:us"
] | automatic-speech-recognition | 2023-07-09T13:43:58Z | ---
language:
- dv
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_13_0
metrics:
- wer
model-index:
- name: Whisper Small Dv
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice 13
type: mozilla-foundation/common_voice_13_0
config: dv
split: test
args: dv
metrics:
- name: Wer
type: wer
value: 110.95037729944013
---
<!-- 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. -->
# Whisper Small Dv
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 13 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1971
- Wer Ortho: 206.4141
- Wer: 110.9504
## 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: 8
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 50
- training_steps: 500
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:--------:|
| 0.1714 | 0.82 | 500 | 0.1971 | 206.4141 | 110.9504 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
hyperr123/mematibas | hyperr123 | 2023-07-09T14:39:57Z | 0 | 0 | null | [
"music",
"tr",
"license:openrail",
"region:us"
] | null | 2023-07-09T13:53:48Z | ---
license: openrail
language:
- tr
tags:
- music
--- |
antolin/distilroberta-base-csn-python-bimodal | antolin | 2023-07-09T14:34:22Z | 86 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"fill-mask",
"dataset:code_search_net",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-07-09T09:24:49Z | ---
datasets:
- code_search_net
widget:
- text: "def <mask> ( a, b ) : if a > b : return a else return b</s>return the maximum value"
- text: "def <mask> ( a, b ) : if a > b : return a else return b"
---
# Model Architecture
This model follows the distilroberta-base architecture. Futhermore, this model was initialized with the checkpoint of distilroberta-base.
# Pre-training phase
This model was pre-trained with the MLM objective (`mlm_probability=0.15`).
During this phase, the inputs had the following format:
$$\left[[CLS], t_1, \dots, t_n, [SEP], w_1, \dots, w_m\right[EOS]]$$
where $t_1, \dots, t_n$ are the code tokens and $w_1, \dots, w_m$ are the natural language description tokens. More concretely, this is the snippet that tokenizes the input:
```python
def tokenize_function_bimodal(examples, tokenizer, max_len):
codes = [' '.join(example) for example in examples['func_code_tokens']]
nls = [' '.join(example) for example in examples['func_documentation_tokens']]
pairs = [[c, nl] for c, nl in zip(codes, nls)]
return tokenizer(pairs, max_length=max_len, padding="max_length", truncation=True)
```
# Training details
- Max length: 512
- Effective batch size: 64
- Total steps: 60000
- Learning rate: 5e-4
# Usage
```python
model = AutoModelForMaskedLM.from_pretrained('antolin/distilroberta-base-csn-python-bimodal')
tokenizer = AutoTokenizer.from_pretrained('antolin/distilroberta-base-csn-python-bimodal')
mask_filler = pipeline("fill-mask", model=model, tokenizer=tokenizer)
code_tokens = ["def", "<mask>", "(", "a", ",", "b", ")", ":", "if", "a", ">", "b", ":", "return", "a", "else", "return", "b"]
nl_tokens = ["return", "the", "maximum", "value"]
input_text = ' '.join(code_tokens) + tokenizer.sep_token + ' '.join(nl_tokens)
pprint(mask_filler(input_text, top_k=5))
```
```shell
[{'score': 0.4645618796348572,
'sequence': 'def max ( a, b ) : if a > b : return a else return b return '
'the maximum value',
'token': 19220,
'token_str': ' max'},
{'score': 0.40963634848594666,
'sequence': 'def maximum ( a, b ) : if a > b : return a else return b '
'return the maximum value',
'token': 4532,
'token_str': ' maximum'},
{'score': 0.02103462442755699,
'sequence': 'def min ( a, b ) : if a > b : return a else return b return '
'the maximum value',
'token': 5251,
'token_str': ' min'},
{'score': 0.014217409305274487,
'sequence': 'def value ( a, b ) : if a > b : return a else return b return '
'the maximum value',
'token': 923,
'token_str': ' value'},
{'score': 0.010762304067611694,
'sequence': 'def minimum ( a, b ) : if a > b : return a else return b '
'return the maximum value',
'token': 3527,
'token_str': ' minimum'}]
```
|
LarryAIDraw/chiyotest | LarryAIDraw | 2023-07-09T14:32:21Z | 0 | 0 | null | [
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-07-09T14:27:32Z | ---
license: creativeml-openrail-m
---
https://civitai.com/models/105174/chiyo-the-sister-of-the-woods-with-a-thousand-young |
LarryAIDraw/calamiti | LarryAIDraw | 2023-07-09T14:32:09Z | 0 | 0 | null | [
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-07-09T14:27:06Z | ---
license: creativeml-openrail-m
---
https://civitai.com/models/105180/calamity-jane-fate-grand-order |
LarryAIDraw/PopeJoan | LarryAIDraw | 2023-07-09T14:31:56Z | 0 | 0 | null | [
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-07-09T14:26:23Z | ---
license: creativeml-openrail-m
---
https://civitai.com/models/105470/pope-joan-or-fategrand-order |
LarryAIDraw/signora_DG | LarryAIDraw | 2023-07-09T14:31:30Z | 0 | 0 | null | [
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-07-09T14:25:34Z | ---
license: creativeml-openrail-m
---
https://civitai.com/models/105495/signora-genshin-impact |
sd-concepts-library/ethos-spirit | sd-concepts-library | 2023-07-09T14:24:01Z | 3 | 0 | transformers | [
"transformers",
"base_model:stabilityai/stable-diffusion-2",
"base_model:finetune:stabilityai/stable-diffusion-2",
"license:mit",
"endpoints_compatible",
"region:us"
] | null | 2023-07-09T11:56:23Z | ---
license: mit
base_model: stabilityai/stable-diffusion-2
---
### Ethos Spirit on Stable Diffusion
This is the `<ethos-spirit>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as a `style`:
|
hafidikhsan/wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-cut | hafidikhsan | 2023-07-09T14:22:45Z | 78 | 0 | transformers | [
"transformers",
"pytorch",
"wav2vec2",
"audio-classification",
"generated_from_trainer",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | audio-classification | 2023-07-09T14:21:39Z | ---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
- precision
- recall
model-index:
- name: wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-cut
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. -->
# wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-cut
This model is a fine-tuned version of [jonatasgrosman/wav2vec2-large-xlsr-53-english](https://huggingface.co/jonatasgrosman/wav2vec2-large-xlsr-53-english) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4268
- Accuracy: 0.6619
- F1: 0.6587
- Precision: 0.6571
- Recall: 0.6619
## 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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
| 0.8845 | 1.0 | 308 | 0.8101 | 0.6457 | 0.5619 | 0.6263 | 0.6457 |
| 0.6667 | 2.0 | 616 | 0.8533 | 0.6209 | 0.6229 | 0.6477 | 0.6209 |
| 0.5061 | 3.0 | 924 | 0.9222 | 0.6534 | 0.6394 | 0.6375 | 0.6534 |
| 0.375 | 4.0 | 1232 | 1.2778 | 0.6611 | 0.6575 | 0.6570 | 0.6611 |
| 0.0501 | 5.0 | 1540 | 1.4268 | 0.6619 | 0.6587 | 0.6571 | 0.6619 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
NasimB/gpt2-concat-guten-mod-rarity-iorder-e1k-ep1k | NasimB | 2023-07-09T14:10:04Z | 6 | 0 | transformers | [
"transformers",
"pytorch",
"gpt2",
"text-generation",
"generated_from_trainer",
"dataset:generator",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-07-09T08:09:40Z | ---
license: mit
tags:
- generated_from_trainer
datasets:
- generator
model-index:
- name: gpt2-concat-guten-mod-rarity-iorder-e1k-ep1k
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. -->
# gpt2-concat-guten-mod-rarity-iorder-e1k-ep1k
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the generator dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1764
## 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: 0.0005
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 6
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:-----:|:---------------:|
| 6.7142 | 0.29 | 500 | 5.6402 |
| 5.3675 | 0.59 | 1000 | 5.2080 |
| 5.0162 | 0.88 | 1500 | 4.9495 |
| 4.7415 | 1.17 | 2000 | 4.7972 |
| 4.5796 | 1.47 | 2500 | 4.6741 |
| 4.4755 | 1.76 | 3000 | 4.5646 |
| 4.3352 | 2.05 | 3500 | 4.4846 |
| 4.1424 | 2.35 | 4000 | 4.4381 |
| 4.1237 | 2.64 | 4500 | 4.3785 |
| 4.0767 | 2.93 | 5000 | 4.3254 |
| 3.866 | 3.23 | 5500 | 4.3222 |
| 3.8116 | 3.52 | 6000 | 4.2890 |
| 3.8011 | 3.81 | 6500 | 4.2546 |
| 3.6876 | 4.11 | 7000 | 4.2563 |
| 3.5206 | 4.4 | 7500 | 4.2498 |
| 3.5211 | 4.69 | 8000 | 4.2322 |
| 3.5138 | 4.99 | 8500 | 4.2213 |
| 3.3361 | 5.28 | 9000 | 4.2348 |
| 3.3334 | 5.57 | 9500 | 4.2340 |
| 3.3228 | 5.87 | 10000 | 4.2330 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
|
k1101jh/ppo-Huggy | k1101jh | 2023-07-09T14:02:33Z | 0 | 0 | ml-agents | [
"ml-agents",
"tensorboard",
"onnx",
"Huggy",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-Huggy",
"region:us"
] | reinforcement-learning | 2023-07-09T14:02:29Z | ---
library_name: ml-agents
tags:
- Huggy
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-Huggy
---
# **ppo** Agent playing **Huggy**
This is a trained model of a **ppo** agent playing **Huggy**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: k1101jh/ppo-Huggy
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
RogerB/KinyaBERT-small-finetuned-kintweetsD | RogerB | 2023-07-09T13:57:42Z | 94 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"bert",
"fill-mask",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-07-09T13:50:29Z | ---
tags:
- generated_from_trainer
model-index:
- name: KinyaBERT-small-finetuned-kintweetsD
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. -->
# KinyaBERT-small-finetuned-kintweetsD
This model is a fine-tuned version of [jean-paul/KinyaBERT-small](https://huggingface.co/jean-paul/KinyaBERT-small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.8590
## 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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 4.339 | 1.0 | 900 | 3.9584 |
| 4.0319 | 2.0 | 1800 | 3.8580 |
| 3.924 | 3.0 | 2700 | 3.8051 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
NasimB/gpt2-concat-mod-datatsets-rarity-all-iorder-no-cut | NasimB | 2023-07-09T13:56:02Z | 18 | 0 | transformers | [
"transformers",
"pytorch",
"gpt2",
"text-generation",
"generated_from_trainer",
"dataset:generator",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-07-09T12:04:45Z | ---
license: mit
tags:
- generated_from_trainer
datasets:
- generator
model-index:
- name: gpt2-concat-mod-datatsets-rarity-all-iorder-no-cut
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. -->
# gpt2-concat-mod-datatsets-rarity-all-iorder-no-cut
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the generator dataset.
It achieves the following results on the evaluation set:
- Loss: 3.2939
## 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: 0.0005
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 6
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 6.7925 | 0.32 | 500 | 5.7247 |
| 5.4528 | 0.64 | 1000 | 5.2926 |
| 5.1083 | 0.96 | 1500 | 5.0277 |
| 4.8149 | 1.28 | 2000 | 4.8716 |
| 4.6803 | 1.6 | 2500 | 4.7459 |
| 4.5724 | 1.93 | 3000 | 4.6411 |
| 4.3471 | 2.25 | 3500 | 4.5849 |
| 4.2732 | 2.57 | 4000 | 4.5154 |
| 4.2272 | 2.89 | 4500 | 4.4580 |
| 4.0295 | 3.21 | 5000 | 4.4423 |
| 3.948 | 3.53 | 5500 | 4.4057 |
| 3.9259 | 3.85 | 6000 | 4.3781 |
| 3.7712 | 4.17 | 6500 | 4.3799 |
| 3.6471 | 4.49 | 7000 | 4.3703 |
| 3.6421 | 4.81 | 7500 | 4.3553 |
| 3.5545 | 5.13 | 8000 | 4.3638 |
| 3.447 | 5.46 | 8500 | 4.3651 |
| 3.4478 | 5.78 | 9000 | 4.3656 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
|
nitishkumargundapu793/chat-bot_response | nitishkumargundapu793 | 2023-07-09T13:26:03Z | 0 | 0 | null | [
"region:us"
] | null | 2023-07-09T13:11:03Z | ---
title: Chat Bot Response
emoji: 👁
colorFrom: indigo
colorTo: yellow
sdk: gradio
sdk_version: 3.0.11
app_file: app.py
pinned: false
---
Check out the configuration reference at https://huggingface.co/docs/hub/spaces#reference
|
rlatmddus159/naval | rlatmddus159 | 2023-07-09T13:15:11Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"safetensors",
"bert",
"fill-mask",
"korean",
"klue",
"ko",
"arxiv:2105.09680",
"arxiv:1910.09700",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | fill-mask | 2023-06-30T15:51:09Z | ---
language: ko
license: cc-by-sa-4.0
tags:
- korean
- klue
mask_token: "[MASK]"
widget:
- text: 대한민국의 수도는 [MASK] 입니다.
---
# KLUE BERT base
## Table of Contents
- [Model Details](#model-details)
- [How to Get Started With the Model](#how-to-get-started-with-the-model)
- [Uses](#uses)
- [Risks, Limitations and Biases](#risks-limitations-and-biases)
- [Training](#training)
- [Evaluation](#evaluation)
- [Environmental Impact](#environmental-impact)
- [Technical Specifications](#technical-specifications)
- [Citation Information](#citation-information)
- [Model Card Authors](#model-card-authors)
## Model Details
**Model Description:** KLUE BERT base is a pre-trained BERT Model on Korean Language. The developers of KLUE BERT base developed the model in the context of the development of the [Korean Language Understanding Evaluation (KLUE) Benchmark](https://arxiv.org/pdf/2105.09680.pdf).
- **Developed by:** See [GitHub Repo](https://github.com/facebookresearch/fairseq/tree/main/examples/roberta) for model developers
- **Model Type:** Transformer-based language model
- **Language(s):** Korean
- **License:** cc-by-sa-4.0
- **Parent Model:** See the [BERT base uncased model](https://huggingface.co/bert-base-uncased) for more information about the BERT base model.
- **Resources for more information:**
- [Research Paper](https://arxiv.org/abs/2105.09680)
- [GitHub Repo](https://github.com/KLUE-benchmark/KLUE)
## How to Get Started With the Model
```python
from transformers import AutoModel, AutoTokenizer
model = AutoModel.from_pretrained("klue/bert-base")
tokenizer = AutoTokenizer.from_pretrained("klue/bert-base")
```
## Uses
#### Direct Use
The model can be used for tasks including topic classification, semantic textual similarity, natural language inference, named entity recognition, and other tasks outlined in the [KLUE Benchmark](https://github.com/KLUE-benchmark/KLUE).
#### Misuse and Out-of-scope Use
The model should not be used to intentionally create hostile or alienating environments for people. In addition, the model was not trained to be factual or true representations of people or events, and therefore using the model to generate such content is out-of-scope for the abilities of this model.
## Risks, Limitations and Biases
Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)). The model developers discuss several ethical considerations related to the model in the [paper](https://arxiv.org/pdf/2105.09680.pdf), including:
- Bias issues with the publicly available data used in the pretraining corpora (and considerations related to filtering)
- PII in the data used in the pretraining corpora (and efforts to pseudonymize the data)
For ethical considerations related to the KLUE Benchmark, also see the [paper](https://arxiv.org/pdf/2105.09680.pdf).
## Training
#### Training Data
The authors use the following pretraining corpora for the model, described in the [associated paper](https://arxiv.org/pdf/2105.09680.pdf):
> We gather the following five publicly available Korean corpora from diverse sources to cover a broad set of topics and many different styles. We combine these corpora to build the final pretraining corpus of size approximately 62GB.
>
> - **MODU:** [Modu Corpus](https://corpus.korean.go.kr) is a collection of Korean corpora distributed by [National Institute of Korean Languages](https://corpus.korean.go.kr/). It includes both formal articles (news and books) and colloquial text (dialogues).
> - **CC-100-Kor:** [CC-100](https://data.statmt.org/cc-100/) is the large-scale multilingual web crawled corpora by using CC-Net ([Wenzek et al., 2020](https://www.aclweb.org/anthology/2020.lrec-1.494)). This is used for training XLM-R ([Conneau et al., 2020](https://aclanthology.org/2020.acl-main.747/)). We use the Korean portion from this corpora.
> - **NAMUWIKI:** NAMUWIKI is a Korean web-based encyclopedia, similar to Wikipedia, but known to be less formal. Specifically, we download [the dump](http://dump.thewiki.kr) created on March 2nd, 2020.
> - **NEWSCRAWL:** NEWSCRAWL consists of 12,800,000 news articles published from 2011 to 2020, collected from a news aggregation platform.
> - **PETITION:** Petition is a collection of public petitions posted to the Blue House asking for administrative actions on social issues. We use the articles in the [Blue House National Petition](https://www1.president.go.kr/petitions) published from [August 2017 to March 2019](https://ko-nlp.github.io/Korpora/en-docs/corpuslist/korean_petitions.html).
The authors also describe ethical considerations related to the pretraining corpora in the [associated paper](https://arxiv.org/pdf/2105.09680.pdf).
#### Training Procedure
##### Preprocessing
The authors describe their preprocessing procedure in the [associated paper](https://arxiv.org/pdf/2105.09680.pdf):
> We filter noisy text and non-Korean text using the same methods from Section 2.3 (of the paper). Each document in the corpus is split into sentences using C++ implementation (v1.3.1.) of rule-based [Korean Sentence Splitter (KSS)](https://github.com/likejazz/korean-sentence-splitter). For CC-100-Kor and NEWSCRAWL, we keep sentences of length greater than equal to 200 characters, as a heuristics to keep well-formed sentences. We then remove sentences included in our benchmark task datasets, using BM25 as a sentence similarity metric ([reference](https://www.microsoft.com/en-us/research/publication/okapi-at-trec-3/)).
###### Tokenization
The authors describe their tokenization procedure in the [associated paper](https://arxiv.org/pdf/2105.09680.pdf):
> We design and use a new tokenization method, morpheme-based subword tokenization. When building a vocabulary, we pre-tokenize a raw text into morphemes using a morphological analyzer, and then we apply byte pair encoding (BPE) ([Senrich et al., 2016](https://aclanthology.org/P16-1162/)) to get the final vocabulary. For morpheme segmentation, we use [Mecab-ko](https://bitbucket.org/eunjeon/mecab-ko), MeCab ([Kudo, 2006](https://taku910.github.io/mecab/)) adapted for Korean, and for BPE segmentation, we use the wordpiece tokenizer from [Huggingface Tokenizers library](https://github.com/huggingface/tokenizers). We specify the vocabulary size to 32k. After building the vocabulary, we only use the BPE model during inference, which allows us to tokenize a word sequence by reflecting morphemes without a morphological analyzer. This improves both usability and speed.
The training configurations are further described in the [paper](https://arxiv.org/pdf/2105.09680.pdf).
## Evaluation
#### Testing Data, Factors and Metrics
The model was evaluated on the [KLUE Benchmark](https://github.com/KLUE-benchmark/KLUE). The tasks and metrics from the KLUE Benchmark that were used to evaluate this model are described briefly below. For more information about the KLUE Benchmark, see the [data card](https://huggingface.co/datasets/klue), [Github Repository](https://github.com/KLUE-benchmark/KLUE), and [associated paper](https://arxiv.org/pdf/2105.09680.pdf).
- **Task:** Topic Classification (TC) - Yonhap News Agency Topic Classification (YNAT), **Metrics:** Macro F1 score, defined as the mean of topic-wise F1 scores, giving the same importance to each topic.
- **Task:** Semantic Textual Similarity (STS), **Metrics:** Pearsons' correlation coefficient (Pearson’ r) and F1 score
- **Task:** Natural Language Inference (NLI), **Metrics:** Accuracy
- **Task:** Named Entity Recognition (NER), **Metrics:** Entity-level macro F1 (Entity F1) and character-level macro F1 (Char F1) scores
- **Task:** Relation Extraction (RE), **Metrics:** Micro F1 score on relation existing cases and area under the precision- recall curve (AUPRC) on all classes
- **Task:** Dependency Parsing (DP), **Metrics:** Unlabeled attachment score (UAS) and labeled attachment score (LAS)
- **Task:** Machine Reading Comprehension (MRC), **Metrics:** Exact match (EM) and character-level ROUGE-W (ROUGE), which can be viewed as longest common consecutive subsequence (LCCS)-based F1 score.
- **Task:** Dialogue State Tracking (DST), **Metrics:** Joint goal accuracy (JGA) and slot micro F1 score (Slot F1)
#### Results
| Task | TC | STS | | NLI | NER | | RE | | DP | | MRC | | DST | |
| :---: |:---: | :---: | :---: |:---:| :---: | :---: |:---:| :---:| :---: |:---: | :---: | :---:| :---: | :---: |
| Metric | F1 | Pearsons' r| F1 | ACC | Entity F1 | Char F1 | F1 | AUPRC| UAS | LAS | EM | ROUGE| JGA |Slot F1 |
| | 85.73| 90.85 | 82.84 |81.63| 83.97 | 91.39 |66.44| 66.17| 89.96 |88.05 | 62.32 | 68.51| 46.64 | 91.61 |
## Environmental Impact
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). We present the hardware type based on the [associated paper](https://arxiv.org/pdf/2105.09680.pdf).
- **Hardware Type:** TPU v3-8
- **Hours used:** Unknown
- **Cloud Provider:** Unknown
- **Compute Region:** Unknown
- **Carbon Emitted:** Unknown
## Technical Specifications
See the [associated paper](https://arxiv.org/pdf/2105.09680.pdf) for details on the modeling architecture (BERT), objective, compute infrastructure, and training details.
## Citation Information
```bibtex
@misc{park2021klue,
title={KLUE: Korean Language Understanding Evaluation},
author={Sungjoon Park and Jihyung Moon and Sungdong Kim and Won Ik Cho and Jiyoon Han and Jangwon Park and Chisung Song and Junseong Kim and Yongsook Song and Taehwan Oh and Joohong Lee and Juhyun Oh and Sungwon Lyu and Younghoon Jeong and Inkwon Lee and Sangwoo Seo and Dongjun Lee and Hyunwoo Kim and Myeonghwa Lee and Seongbo Jang and Seungwon Do and Sunkyoung Kim and Kyungtae Lim and Jongwon Lee and Kyumin Park and Jamin Shin and Seonghyun Kim and Lucy Park and Alice Oh and Jungwoo Ha and Kyunghyun Cho},
year={2021},
eprint={2105.09680},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|
WALIDALI/cynthiily | WALIDALI | 2023-07-09T12:43:28Z | 2 | 0 | diffusers | [
"diffusers",
"safetensors",
"text-to-image",
"stable-diffusion",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] | text-to-image | 2023-07-09T12:39:33Z | ---
license: creativeml-openrail-m
tags:
- text-to-image
- stable-diffusion
---
### cynthiily Dreambooth model trained by WALIDALI with [TheLastBen's fast-DreamBooth](https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast-DreamBooth.ipynb) notebook
Test the concept via A1111 Colab [fast-Colab-A1111](https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast_stable_diffusion_AUTOMATIC1111.ipynb)
Sample pictures of this concept:
|
hafidikhsan/wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-cut-balance | hafidikhsan | 2023-07-09T12:33:51Z | 88 | 5 | transformers | [
"transformers",
"pytorch",
"wav2vec2",
"audio-classification",
"generated_from_trainer",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | audio-classification | 2023-07-09T12:31:36Z | ---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
- precision
- recall
model-index:
- name: wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-cut-balance
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. -->
# wav2vec2-large-xlsr-53-english-pronunciation-evaluation-aod-cut-balance
This model is a fine-tuned version of [jonatasgrosman/wav2vec2-large-xlsr-53-english](https://huggingface.co/jonatasgrosman/wav2vec2-large-xlsr-53-english) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0674
- Accuracy: 0.6055
- F1: 0.6017
- Precision: 0.6074
- Recall: 0.6055
## 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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
| 1.0011 | 1.0 | 105 | 1.0494 | 0.5 | 0.4111 | 0.4721 | 0.5 |
| 0.7777 | 2.0 | 210 | 0.9454 | 0.5576 | 0.5178 | 0.5332 | 0.5576 |
| 0.7462 | 3.0 | 315 | 1.1190 | 0.5815 | 0.5649 | 0.5757 | 0.5815 |
| 0.6099 | 4.0 | 420 | 1.0299 | 0.6043 | 0.5975 | 0.5992 | 0.6043 |
| 0.4457 | 5.0 | 525 | 1.0674 | 0.6055 | 0.6017 | 0.6074 | 0.6055 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
yufeih/models | yufeih | 2023-07-09T12:33:16Z | 0 | 0 | null | [
"onnx",
"region:us"
] | null | 2023-07-09T12:31:02Z | # Models
## private-detector.onnx
An image classifier that detect lewd images: https://github.com/bumble-tech/private-detector:
1. Download the saved model here: https://storage.googleapis.com/private_detector/private_detector_with_frozen.zip
2. Install [tensorflow-onnx](https://github.com/onnx/tensorflow-onnx)
3. Run `python -m tf2onnx.convert --saved-model model/saved_model --output private-detector.onnx`
## toxic-comment
This model is a fine-tuned version of the DistilBERT model to classify toxic comments: https://huggingface.co/martin-ha/toxic-comment-model
1. Run `optimum-cli export onnx --model martin-ha/toxic-comment-model toxic-comment/` |
carybr/ppo-LunarLander-v2 | carybr | 2023-07-09T12:29:21Z | 0 | 0 | stable-baselines3 | [
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] | reinforcement-learning | 2023-07-09T12:27:12Z | ---
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: 262.36 +/- 19.76
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
...
```
|
robsucher/opt-6.7b-lora-tagger-english-quotes | robsucher | 2023-07-09T12:22:53Z | 1 | 0 | peft | [
"peft",
"region:us"
] | null | 2023-07-09T12:22:52Z | ---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: True
- load_in_4bit: False
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.4.0.dev0
|
jordyvl/dit-tiny_tobacco3482_kd_CEKD_t5.0_a0.9 | jordyvl | 2023-07-09T12:17:46Z | 134 | 0 | transformers | [
"transformers",
"pytorch",
"beit",
"image-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | image-classification | 2023-07-09T12:01:43Z | ---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: dit-tiny_tobacco3482_kd_CEKD_t5.0_a0.9
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. -->
# dit-tiny_tobacco3482_kd_CEKD_t5.0_a0.9
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.5147
- Accuracy: 0.18
- Brier Loss: 0.8746
- Nll: 6.7241
- F1 Micro: 0.18
- F1 Macro: 0.0306
- Ece: 0.2451
- Aurc: 0.8494
## 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: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:-------:|:--------:|:--------:|:------:|:------:|
| No log | 0.96 | 3 | 2.6571 | 0.145 | 0.8999 | 10.1542 | 0.145 | 0.0253 | 0.2220 | 0.8466 |
| No log | 1.96 | 6 | 2.6281 | 0.145 | 0.8947 | 10.5635 | 0.145 | 0.0253 | 0.2236 | 0.8461 |
| No log | 2.96 | 9 | 2.5865 | 0.14 | 0.8870 | 8.5822 | 0.14 | 0.0433 | 0.2063 | 0.8040 |
| No log | 3.96 | 12 | 2.5552 | 0.19 | 0.8811 | 6.5445 | 0.19 | 0.0552 | 0.2421 | 0.8576 |
| No log | 4.96 | 15 | 2.5387 | 0.155 | 0.8782 | 7.1184 | 0.155 | 0.0277 | 0.2280 | 0.8892 |
| No log | 5.96 | 18 | 2.5317 | 0.18 | 0.8774 | 8.7285 | 0.18 | 0.0319 | 0.2392 | 0.8538 |
| No log | 6.96 | 21 | 2.5274 | 0.18 | 0.8770 | 8.2533 | 0.18 | 0.0306 | 0.2476 | 0.8514 |
| No log | 7.96 | 24 | 2.5238 | 0.18 | 0.8767 | 6.9903 | 0.18 | 0.0306 | 0.2465 | 0.8523 |
| No log | 8.96 | 27 | 2.5205 | 0.18 | 0.8762 | 6.9049 | 0.18 | 0.0306 | 0.2473 | 0.8528 |
| No log | 9.96 | 30 | 2.5189 | 0.18 | 0.8758 | 6.8830 | 0.18 | 0.0306 | 0.2515 | 0.8526 |
| No log | 10.96 | 33 | 2.5180 | 0.18 | 0.8756 | 6.8133 | 0.18 | 0.0306 | 0.2469 | 0.8522 |
| No log | 11.96 | 36 | 2.5175 | 0.18 | 0.8754 | 6.7422 | 0.18 | 0.0306 | 0.2500 | 0.8519 |
| No log | 12.96 | 39 | 2.5173 | 0.18 | 0.8753 | 6.5762 | 0.18 | 0.0306 | 0.2533 | 0.8515 |
| No log | 13.96 | 42 | 2.5168 | 0.18 | 0.8751 | 6.5666 | 0.18 | 0.0306 | 0.2528 | 0.8516 |
| No log | 14.96 | 45 | 2.5164 | 0.18 | 0.8750 | 6.7246 | 0.18 | 0.0306 | 0.2532 | 0.8512 |
| No log | 15.96 | 48 | 2.5160 | 0.18 | 0.8750 | 6.7221 | 0.18 | 0.0306 | 0.2456 | 0.8507 |
| No log | 16.96 | 51 | 2.5157 | 0.18 | 0.8749 | 6.7242 | 0.18 | 0.0306 | 0.2457 | 0.8507 |
| No log | 17.96 | 54 | 2.5158 | 0.18 | 0.8749 | 6.7241 | 0.18 | 0.0306 | 0.2417 | 0.8503 |
| No log | 18.96 | 57 | 2.5157 | 0.18 | 0.8749 | 6.7259 | 0.18 | 0.0306 | 0.2455 | 0.8503 |
| No log | 19.96 | 60 | 2.5153 | 0.18 | 0.8748 | 6.7248 | 0.18 | 0.0306 | 0.2452 | 0.8495 |
| No log | 20.96 | 63 | 2.5151 | 0.18 | 0.8748 | 6.7250 | 0.18 | 0.0306 | 0.2414 | 0.8494 |
| No log | 21.96 | 66 | 2.5149 | 0.18 | 0.8747 | 6.7250 | 0.18 | 0.0306 | 0.2452 | 0.8495 |
| No log | 22.96 | 69 | 2.5147 | 0.18 | 0.8747 | 6.7247 | 0.18 | 0.0306 | 0.2451 | 0.8495 |
| No log | 23.96 | 72 | 2.5147 | 0.18 | 0.8747 | 6.7246 | 0.18 | 0.0306 | 0.2451 | 0.8495 |
| No log | 24.96 | 75 | 2.5147 | 0.18 | 0.8746 | 6.7241 | 0.18 | 0.0306 | 0.2451 | 0.8494 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
|
Serjssv/distilhubert-finetuned-gtzan | Serjssv | 2023-07-09T12:14:15Z | 129 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"hubert",
"audio-classification",
"generated_from_trainer",
"dataset:marsyas/gtzan",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | audio-classification | 2023-07-05T06:56:21Z | ---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- marsyas/gtzan
metrics:
- accuracy
model-index:
- name: distilhubert-finetuned-gtzan-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. -->
# distilhubert-finetuned-gtzan-1
This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5778
- Accuracy: 0.82
## 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: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 14
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 2.103 | 1.0 | 112 | 2.1288 | 0.42 |
| 1.5948 | 2.0 | 225 | 1.6203 | 0.55 |
| 1.3883 | 3.0 | 337 | 1.2437 | 0.69 |
| 1.1032 | 4.0 | 450 | 1.0490 | 0.73 |
| 0.7595 | 5.0 | 562 | 0.8857 | 0.79 |
| 0.812 | 6.0 | 675 | 0.7776 | 0.8 |
| 0.4903 | 7.0 | 787 | 0.7682 | 0.78 |
| 0.5568 | 8.0 | 900 | 0.7100 | 0.79 |
| 0.405 | 9.0 | 1012 | 0.6279 | 0.84 |
| 0.5888 | 10.0 | 1125 | 0.6944 | 0.8 |
| 0.2576 | 11.0 | 1237 | 0.6027 | 0.83 |
| 0.2123 | 12.0 | 1350 | 0.5891 | 0.83 |
| 0.2008 | 13.0 | 1462 | 0.5659 | 0.83 |
| 0.1343 | 13.94 | 1568 | 0.5778 | 0.82 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
fmod99/bert-finetuned-ner | fmod99 | 2023-07-09T12:10:12Z | 85 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"bert",
"token-classification",
"generated_from_trainer",
"dataset:conll2003",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | token-classification | 2023-07-01T19:15:41Z | ---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- conll2003
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: bert-finetuned-ner
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: conll2003
type: conll2003
config: conll2003
split: validation
args: conll2003
metrics:
- name: Precision
type: precision
value: 0.9353184449958644
- name: Recall
type: recall
value: 0.9515314708852238
- name: F1
type: f1
value: 0.9433553015767081
- name: Accuracy
type: accuracy
value: 0.9867840113027609
---
<!-- 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. -->
# bert-finetuned-ner
This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on the conll2003 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0618
- Precision: 0.9353
- Recall: 0.9515
- F1: 0.9434
- Accuracy: 0.9868
## 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: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| 0.0854 | 1.0 | 1756 | 0.0706 | 0.9143 | 0.9318 | 0.9230 | 0.9817 |
| 0.0332 | 2.0 | 3512 | 0.0648 | 0.9310 | 0.9498 | 0.9404 | 0.9862 |
| 0.017 | 3.0 | 5268 | 0.0618 | 0.9353 | 0.9515 | 0.9434 | 0.9868 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
jordyvl/dit-small_tobacco3482_kd_CEKD_t5.0_a0.7 | jordyvl | 2023-07-09T12:01:00Z | 128 | 0 | transformers | [
"transformers",
"pytorch",
"beit",
"image-classification",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | image-classification | 2023-07-09T11:43:00Z | ---
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: dit-small_tobacco3482_kd_CEKD_t5.0_a0.7
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. -->
# dit-small_tobacco3482_kd_CEKD_t5.0_a0.7
This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1347
- Accuracy: 0.185
- Brier Loss: 0.8666
- Nll: 5.9997
- F1 Micro: 0.185
- F1 Macro: 0.0488
- Ece: 0.2480
- Aurc: 0.7353
## 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: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:------:|:--------:|:--------:|:------:|:------:|
| No log | 0.96 | 3 | 3.3695 | 0.06 | 0.9042 | 9.1505 | 0.06 | 0.0114 | 0.1750 | 0.9033 |
| No log | 1.96 | 6 | 3.2847 | 0.18 | 0.8890 | 7.1646 | 0.18 | 0.0305 | 0.2263 | 0.8027 |
| No log | 2.96 | 9 | 3.2039 | 0.18 | 0.8773 | 8.6118 | 0.18 | 0.0305 | 0.2478 | 0.8186 |
| No log | 3.96 | 12 | 3.1950 | 0.18 | 0.8806 | 7.4891 | 0.18 | 0.0305 | 0.2514 | 0.8131 |
| No log | 4.96 | 15 | 3.1951 | 0.185 | 0.8795 | 6.7125 | 0.185 | 0.0488 | 0.2555 | 0.7835 |
| No log | 5.96 | 18 | 3.1931 | 0.185 | 0.8766 | 5.2600 | 0.185 | 0.0488 | 0.2526 | 0.7702 |
| No log | 6.96 | 21 | 3.1876 | 0.185 | 0.8741 | 5.6453 | 0.185 | 0.0488 | 0.2372 | 0.7672 |
| No log | 7.96 | 24 | 3.1800 | 0.185 | 0.8726 | 5.9473 | 0.185 | 0.0488 | 0.2412 | 0.7644 |
| No log | 8.96 | 27 | 3.1712 | 0.185 | 0.8712 | 5.9421 | 0.185 | 0.0488 | 0.2491 | 0.7615 |
| No log | 9.96 | 30 | 3.1656 | 0.185 | 0.8704 | 6.6276 | 0.185 | 0.0488 | 0.2516 | 0.7602 |
| No log | 10.96 | 33 | 3.1623 | 0.185 | 0.8704 | 6.8796 | 0.185 | 0.0488 | 0.2487 | 0.7598 |
| No log | 11.96 | 36 | 3.1601 | 0.185 | 0.8708 | 7.1352 | 0.185 | 0.0488 | 0.2451 | 0.7559 |
| No log | 12.96 | 39 | 3.1573 | 0.185 | 0.8706 | 7.0151 | 0.185 | 0.0488 | 0.2492 | 0.7531 |
| No log | 13.96 | 42 | 3.1531 | 0.185 | 0.8699 | 6.7912 | 0.185 | 0.0488 | 0.2450 | 0.7484 |
| No log | 14.96 | 45 | 3.1485 | 0.185 | 0.8693 | 6.6578 | 0.185 | 0.0488 | 0.2513 | 0.7491 |
| No log | 15.96 | 48 | 3.1449 | 0.185 | 0.8685 | 6.1407 | 0.185 | 0.0488 | 0.2596 | 0.7463 |
| No log | 16.96 | 51 | 3.1428 | 0.185 | 0.8681 | 5.9160 | 0.185 | 0.0488 | 0.2548 | 0.7432 |
| No log | 17.96 | 54 | 3.1421 | 0.185 | 0.8678 | 5.8419 | 0.185 | 0.0488 | 0.2449 | 0.7401 |
| No log | 18.96 | 57 | 3.1413 | 0.185 | 0.8677 | 5.7417 | 0.185 | 0.0488 | 0.2606 | 0.7382 |
| No log | 19.96 | 60 | 3.1391 | 0.185 | 0.8673 | 5.7824 | 0.185 | 0.0488 | 0.2432 | 0.7365 |
| No log | 20.96 | 63 | 3.1378 | 0.185 | 0.8671 | 5.9509 | 0.185 | 0.0488 | 0.2598 | 0.7368 |
| No log | 21.96 | 66 | 3.1364 | 0.185 | 0.8668 | 6.0164 | 0.185 | 0.0488 | 0.2477 | 0.7361 |
| No log | 22.96 | 69 | 3.1355 | 0.185 | 0.8667 | 6.0109 | 0.185 | 0.0488 | 0.2437 | 0.7352 |
| No log | 23.96 | 72 | 3.1350 | 0.185 | 0.8666 | 6.0029 | 0.185 | 0.0488 | 0.2438 | 0.7351 |
| No log | 24.96 | 75 | 3.1347 | 0.185 | 0.8666 | 5.9997 | 0.185 | 0.0488 | 0.2480 | 0.7353 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
|
MichelNivard/Replit3b_rchat_v5 | MichelNivard | 2023-07-09T12:00:55Z | 6 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"mpt",
"text-generation",
"custom_code",
"dataset:bigcode/the-stack",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | text-generation | 2023-07-08T18:50:51Z | ---
datasets:
- bigcode/the-stack
---
# hexcoder 3b (replit)
This is a model that trains the base [replit code model](https://huggingface.co/replit/replit-code-v1-3b) on R coding instructions, which are brief questions to which I have formulated a brief anwser.
Then I intend to instruction tune the model on all stackoverflow questions and anwsers with the tag 'r' in the 2011 to 2016 timeframe, presenting stackoverflow questions as <|user|> and the best answer as <|assistant|>. This will teach the model that it is expected to produce an answer to a user's question about 'r'.
The intended outcome is a reasonably adequate model which can answer basic r user questions, but more broadly an evaluaion of the data/sources and training needed to produce great open source code generating models for r. |
vioannidis/wav2vec2-large-xls-r-300m-turkish-colab | vioannidis | 2023-07-09T11:53:46Z | 86 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"wav2vec2",
"automatic-speech-recognition",
"generated_from_trainer",
"dataset:common_voice",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | automatic-speech-recognition | 2023-07-09T11:09:25Z | ---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- common_voice
model-index:
- name: wav2vec2-large-xls-r-300m-turkish-colab
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. -->
# wav2vec2-large-xls-r-300m-turkish-colab
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice 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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 3
### Training results
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
|
Subsets and Splits