Search is not available for this dataset
pipeline_tag
stringclasses 48
values | library_name
stringclasses 205
values | text
stringlengths 0
18.3M
| metadata
stringlengths 2
1.07B
| id
stringlengths 5
122
| last_modified
null | tags
sequencelengths 1
1.84k
| sha
null | created_at
stringlengths 25
25
|
|---|---|---|---|---|---|---|---|---|
text-classification | transformers | {} | sc-ai/snippet-reranker-v1 | null | [
"transformers",
"safetensors",
"xlm-roberta",
"text-classification",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:00:27+00:00 |
|
null | null | {} | bbcgjb/chatgpt | null | [
"region:us"
] | null | 2024-04-30T03:01:15+00:00 |
|
null | null | {"license": "apache-2.0"} | luciusy/two_stage_no_prompt | null | [
"license:apache-2.0",
"region:us"
] | null | 2024-04-30T03:03:04+00:00 |
|
null | transformers |
<p align="center">
<img src="https://doctr-static.mindee.com/models?id=v0.3.1/Logo_doctr.gif&src=0" width="60%">
</p>
**Optical Character Recognition made seamless & accessible to anyone, powered by TensorFlow 2 & PyTorch**
## Task: detection
https://github.com/mindee/doctr
### Example usage:
```python
>>> from doctr.io import DocumentFile
>>> from doctr.models import ocr_predictor, from_hub
>>> img = DocumentFile.from_images(['<image_path>'])
>>> # Load your model from the hub
>>> model = from_hub('mindee/my-model')
>>> # Pass it to the predictor
>>> # If your model is a recognition model:
>>> predictor = ocr_predictor(det_arch='db_mobilenet_v3_large',
>>> reco_arch=model,
>>> pretrained=True)
>>> # If your model is a detection model:
>>> predictor = ocr_predictor(det_arch=model,
>>> reco_arch='crnn_mobilenet_v3_small',
>>> pretrained=True)
>>> # Get your predictions
>>> res = predictor(img)
```
### Run Configuration
{
"train_path": "/workspace/donut_train/doctr/train/",
"val_path": "/workspace/donut_train/doctr/val/",
"arch": "db_resnet50",
"name": null,
"epochs": 5,
"batch_size": 10,
"device": 0,
"save_interval_epoch": false,
"input_size": 1024,
"lr": 0.001,
"weight_decay": 0,
"workers": 16,
"resume": null,
"test_only": false,
"freeze_backbone": false,
"show_samples": false,
"wb": false,
"push_to_hub": true,
"pretrained": false,
"rotation": false,
"eval_straight": false,
"sched": "poly",
"amp": false,
"find_lr": false,
"early_stop": false,
"early_stop_epochs": 5,
"early_stop_delta": 0.01
} | {"language": "en"} | Alexleetw/db_resnet50_20240430-025744 | null | [
"transformers",
"pytorch",
"en",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:03:11+00:00 |
null | null | {} | ivykopal/german_adapter_100k | null | [
"region:us"
] | null | 2024-04-30T03:03:33+00:00 |
|
null | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | IN4/fast-whisper-v3-LoRA-8bit-epochs-3_num1_ru_kz | null | [
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:04:23+00:00 |
text-generation | transformers | {} | KYLN24/tmp-Llama-2-7b-chat-hf-webshop-sft | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"conversational",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:05:13+00:00 |
|
null | peft |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
### Framework versions
- PEFT 0.10.1.dev0 | {"library_name": "peft", "base_model": "baffo32/decapoda-research-llama-7B-hf"} | Yuki20/alpaca7B-lora | null | [
"peft",
"safetensors",
"arxiv:1910.09700",
"base_model:baffo32/decapoda-research-llama-7B-hf",
"region:us"
] | null | 2024-04-30T03:05:29+00:00 |
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | lunarsylph/mooncell_v35 | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:06:07+00:00 |
null | peft |
<!-- 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. -->
# GUE_mouse_3-seqsight_16384_512_56M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_mouse_3](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_mouse_3) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9372
- F1 Score: 0.8326
- Accuracy: 0.8326
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:------:|:-----:|:---------------:|:--------:|:--------:|
| 0.5618 | 13.33 | 200 | 0.5233 | 0.7196 | 0.7280 |
| 0.4488 | 26.67 | 400 | 0.4306 | 0.8322 | 0.8326 |
| 0.3592 | 40.0 | 600 | 0.3992 | 0.8368 | 0.8368 |
| 0.3015 | 53.33 | 800 | 0.3865 | 0.8574 | 0.8577 |
| 0.2665 | 66.67 | 1000 | 0.4015 | 0.8444 | 0.8452 |
| 0.2492 | 80.0 | 1200 | 0.3666 | 0.8658 | 0.8661 |
| 0.2203 | 93.33 | 1400 | 0.4055 | 0.8492 | 0.8494 |
| 0.2024 | 106.67 | 1600 | 0.4123 | 0.8615 | 0.8619 |
| 0.1897 | 120.0 | 1800 | 0.4423 | 0.8575 | 0.8577 |
| 0.1705 | 133.33 | 2000 | 0.4844 | 0.8532 | 0.8536 |
| 0.1607 | 146.67 | 2200 | 0.5158 | 0.8614 | 0.8619 |
| 0.1487 | 160.0 | 2400 | 0.4890 | 0.8532 | 0.8536 |
| 0.1393 | 173.33 | 2600 | 0.5045 | 0.8534 | 0.8536 |
| 0.1249 | 186.67 | 2800 | 0.5124 | 0.8619 | 0.8619 |
| 0.1154 | 200.0 | 3000 | 0.5181 | 0.8536 | 0.8536 |
| 0.1125 | 213.33 | 3200 | 0.5425 | 0.8535 | 0.8536 |
| 0.1085 | 226.67 | 3400 | 0.5355 | 0.8702 | 0.8703 |
| 0.0966 | 240.0 | 3600 | 0.5825 | 0.8573 | 0.8577 |
| 0.0938 | 253.33 | 3800 | 0.5689 | 0.8618 | 0.8619 |
| 0.0877 | 266.67 | 4000 | 0.5661 | 0.8618 | 0.8619 |
| 0.0873 | 280.0 | 4200 | 0.5564 | 0.8577 | 0.8577 |
| 0.0786 | 293.33 | 4400 | 0.5884 | 0.8410 | 0.8410 |
| 0.0803 | 306.67 | 4600 | 0.5591 | 0.8619 | 0.8619 |
| 0.0722 | 320.0 | 4800 | 0.5959 | 0.8577 | 0.8577 |
| 0.0741 | 333.33 | 5000 | 0.5965 | 0.8577 | 0.8577 |
| 0.0662 | 346.67 | 5200 | 0.6233 | 0.8577 | 0.8577 |
| 0.0672 | 360.0 | 5400 | 0.6139 | 0.8577 | 0.8577 |
| 0.0623 | 373.33 | 5600 | 0.6260 | 0.8534 | 0.8536 |
| 0.0589 | 386.67 | 5800 | 0.6026 | 0.8494 | 0.8494 |
| 0.0566 | 400.0 | 6000 | 0.6391 | 0.8577 | 0.8577 |
| 0.0554 | 413.33 | 6200 | 0.6336 | 0.8577 | 0.8577 |
| 0.0544 | 426.67 | 6400 | 0.6392 | 0.8577 | 0.8577 |
| 0.0506 | 440.0 | 6600 | 0.6272 | 0.8619 | 0.8619 |
| 0.0488 | 453.33 | 6800 | 0.6549 | 0.8577 | 0.8577 |
| 0.0496 | 466.67 | 7000 | 0.6417 | 0.8661 | 0.8661 |
| 0.0506 | 480.0 | 7200 | 0.6602 | 0.8661 | 0.8661 |
| 0.0452 | 493.33 | 7400 | 0.6630 | 0.8745 | 0.8745 |
| 0.044 | 506.67 | 7600 | 0.6793 | 0.8661 | 0.8661 |
| 0.0432 | 520.0 | 7800 | 0.6656 | 0.8619 | 0.8619 |
| 0.0451 | 533.33 | 8000 | 0.6787 | 0.8619 | 0.8619 |
| 0.0457 | 546.67 | 8200 | 0.6668 | 0.8577 | 0.8577 |
| 0.0415 | 560.0 | 8400 | 0.6715 | 0.8619 | 0.8619 |
| 0.0448 | 573.33 | 8600 | 0.6673 | 0.8619 | 0.8619 |
| 0.0391 | 586.67 | 8800 | 0.6952 | 0.8577 | 0.8577 |
| 0.043 | 600.0 | 9000 | 0.6812 | 0.8619 | 0.8619 |
| 0.039 | 613.33 | 9200 | 0.6920 | 0.8577 | 0.8577 |
| 0.0385 | 626.67 | 9400 | 0.6871 | 0.8619 | 0.8619 |
| 0.04 | 640.0 | 9600 | 0.6837 | 0.8619 | 0.8619 |
| 0.0397 | 653.33 | 9800 | 0.6876 | 0.8619 | 0.8619 |
| 0.0369 | 666.67 | 10000 | 0.6881 | 0.8619 | 0.8619 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_mouse_3-seqsight_16384_512_56M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_mouse_3-seqsight_16384_512_56M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:07:06+00:00 |
text-generation | Transformers |
This is a Llama 2 architecture model series trained on the FineWeb dataset upto 1 Billion parameters and uses tiktoken cl100k_base model as tokenizer | {"license": "mit", "library_name": "Transformers", "datasets": ["HuggingFaceFW/fineweb"], "pipeline_tag": "text-generation"} | sabareesh88/fw14k | null | [
"Transformers",
"text-generation",
"dataset:HuggingFaceFW/fineweb",
"license:mit",
"region:us"
] | null | 2024-04-30T03:07:14+00:00 |
null | null | {} | luciusy/two_stage_batch4_pp2 | null | [
"region:us"
] | null | 2024-04-30T03:07:17+00:00 |
|
null | transformers |
<p align="center">
<img src="https://doctr-static.mindee.com/models?id=v0.3.1/Logo_doctr.gif&src=0" width="60%">
</p>
**Optical Character Recognition made seamless & accessible to anyone, powered by TensorFlow 2 & PyTorch**
## Task: detection
https://github.com/mindee/doctr
### Example usage:
```python
>>> from doctr.io import DocumentFile
>>> from doctr.models import ocr_predictor, from_hub
>>> img = DocumentFile.from_images(['<image_path>'])
>>> # Load your model from the hub
>>> model = from_hub('mindee/my-model')
>>> # Pass it to the predictor
>>> # If your model is a recognition model:
>>> predictor = ocr_predictor(det_arch='db_mobilenet_v3_large',
>>> reco_arch=model,
>>> pretrained=True)
>>> # If your model is a detection model:
>>> predictor = ocr_predictor(det_arch=model,
>>> reco_arch='crnn_mobilenet_v3_small',
>>> pretrained=True)
>>> # Get your predictions
>>> res = predictor(img)
```
### Run Configuration
{
"train_path": "/workspace/donut_train/doctr/train/",
"val_path": "/workspace/donut_train/doctr/val/",
"arch": "db_resnet50",
"name": null,
"epochs": 5,
"batch_size": 2,
"device": 0,
"save_interval_epoch": false,
"input_size": 1024,
"lr": 0.001,
"weight_decay": 0,
"workers": 16,
"resume": null,
"test_only": false,
"freeze_backbone": false,
"show_samples": false,
"wb": false,
"push_to_hub": true,
"pretrained": false,
"rotation": false,
"eval_straight": false,
"sched": "poly",
"amp": false,
"find_lr": false,
"early_stop": false,
"early_stop_epochs": 5,
"early_stop_delta": 0.01
} | {"language": "en"} | Alexleetw/db_resnet50_20240430-030513 | null | [
"transformers",
"pytorch",
"en",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:08:30+00:00 |
null | peft |
<!-- 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. -->
# GUE_mouse_3-seqsight_16384_512_56M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_mouse_3](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_mouse_3) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9635
- F1 Score: 0.8322
- Accuracy: 0.8326
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:------:|:-----:|:---------------:|:--------:|:--------:|
| 0.5109 | 13.33 | 200 | 0.4179 | 0.8322 | 0.8326 |
| 0.3054 | 26.67 | 400 | 0.3763 | 0.8368 | 0.8368 |
| 0.2227 | 40.0 | 600 | 0.4092 | 0.8326 | 0.8326 |
| 0.1641 | 53.33 | 800 | 0.5099 | 0.8449 | 0.8452 |
| 0.1239 | 66.67 | 1000 | 0.6035 | 0.8366 | 0.8368 |
| 0.1032 | 80.0 | 1200 | 0.6306 | 0.8446 | 0.8452 |
| 0.0806 | 93.33 | 1400 | 0.7046 | 0.8282 | 0.8285 |
| 0.0714 | 106.67 | 1600 | 0.7260 | 0.8322 | 0.8326 |
| 0.0579 | 120.0 | 1800 | 0.7198 | 0.8410 | 0.8410 |
| 0.0464 | 133.33 | 2000 | 0.8073 | 0.8744 | 0.8745 |
| 0.0467 | 146.67 | 2200 | 0.9499 | 0.8492 | 0.8494 |
| 0.039 | 160.0 | 2400 | 0.8993 | 0.8491 | 0.8494 |
| 0.0379 | 173.33 | 2600 | 0.8966 | 0.8577 | 0.8577 |
| 0.0321 | 186.67 | 2800 | 0.9570 | 0.8492 | 0.8494 |
| 0.0295 | 200.0 | 3000 | 0.9108 | 0.8535 | 0.8536 |
| 0.0266 | 213.33 | 3200 | 0.9787 | 0.8368 | 0.8368 |
| 0.0288 | 226.67 | 3400 | 0.9061 | 0.8493 | 0.8494 |
| 0.023 | 240.0 | 3600 | 1.0900 | 0.8489 | 0.8494 |
| 0.0228 | 253.33 | 3800 | 0.9683 | 0.8575 | 0.8577 |
| 0.0208 | 266.67 | 4000 | 0.9597 | 0.8492 | 0.8494 |
| 0.02 | 280.0 | 4200 | 1.0382 | 0.8448 | 0.8452 |
| 0.0203 | 293.33 | 4400 | 0.9813 | 0.8534 | 0.8536 |
| 0.0189 | 306.67 | 4600 | 1.2172 | 0.8444 | 0.8452 |
| 0.0151 | 320.0 | 4800 | 1.0190 | 0.8409 | 0.8410 |
| 0.0192 | 333.33 | 5000 | 1.0170 | 0.8368 | 0.8368 |
| 0.0165 | 346.67 | 5200 | 1.0590 | 0.8614 | 0.8619 |
| 0.0125 | 360.0 | 5400 | 1.1316 | 0.8573 | 0.8577 |
| 0.0159 | 373.33 | 5600 | 0.9990 | 0.8493 | 0.8494 |
| 0.0123 | 386.67 | 5800 | 1.0003 | 0.8494 | 0.8494 |
| 0.014 | 400.0 | 6000 | 1.0304 | 0.8575 | 0.8577 |
| 0.0135 | 413.33 | 6200 | 1.0387 | 0.8534 | 0.8536 |
| 0.0115 | 426.67 | 6400 | 1.1494 | 0.8449 | 0.8452 |
| 0.0105 | 440.0 | 6600 | 1.2265 | 0.8573 | 0.8577 |
| 0.0103 | 453.33 | 6800 | 1.1952 | 0.8575 | 0.8577 |
| 0.0083 | 466.67 | 7000 | 1.3091 | 0.8530 | 0.8536 |
| 0.0102 | 480.0 | 7200 | 1.2727 | 0.8406 | 0.8410 |
| 0.0096 | 493.33 | 7400 | 1.1576 | 0.8368 | 0.8368 |
| 0.0098 | 506.67 | 7600 | 1.1458 | 0.8534 | 0.8536 |
| 0.0085 | 520.0 | 7800 | 1.1852 | 0.8534 | 0.8536 |
| 0.0093 | 533.33 | 8000 | 1.1254 | 0.8575 | 0.8577 |
| 0.0103 | 546.67 | 8200 | 1.1174 | 0.8493 | 0.8494 |
| 0.007 | 560.0 | 8400 | 1.1201 | 0.8534 | 0.8536 |
| 0.0098 | 573.33 | 8600 | 1.1218 | 0.8492 | 0.8494 |
| 0.0083 | 586.67 | 8800 | 1.1587 | 0.8573 | 0.8577 |
| 0.0083 | 600.0 | 9000 | 1.1507 | 0.8493 | 0.8494 |
| 0.0093 | 613.33 | 9200 | 1.1114 | 0.8534 | 0.8536 |
| 0.0055 | 626.67 | 9400 | 1.1478 | 0.8493 | 0.8494 |
| 0.0072 | 640.0 | 9600 | 1.1360 | 0.8493 | 0.8494 |
| 0.0075 | 653.33 | 9800 | 1.1402 | 0.8577 | 0.8577 |
| 0.006 | 666.67 | 10000 | 1.1445 | 0.8577 | 0.8577 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_mouse_3-seqsight_16384_512_56M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_mouse_3-seqsight_16384_512_56M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:09:23+00:00 |
null | peft |
<!-- 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. -->
# GUE_mouse_3-seqsight_16384_512_56M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_mouse_3](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_mouse_3) dataset.
It achieves the following results on the evaluation set:
- Loss: 1.3820
- F1 Score: 0.8534
- Accuracy: 0.8536
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:------:|:-----:|:---------------:|:--------:|:--------:|
| 0.4475 | 13.33 | 200 | 0.3527 | 0.8536 | 0.8536 |
| 0.2157 | 26.67 | 400 | 0.4971 | 0.8238 | 0.8243 |
| 0.1214 | 40.0 | 600 | 0.5771 | 0.8323 | 0.8326 |
| 0.0693 | 53.33 | 800 | 0.6868 | 0.8534 | 0.8536 |
| 0.0492 | 66.67 | 1000 | 0.7634 | 0.8575 | 0.8577 |
| 0.0428 | 80.0 | 1200 | 0.8284 | 0.8492 | 0.8494 |
| 0.0346 | 93.33 | 1400 | 0.8598 | 0.8533 | 0.8536 |
| 0.0259 | 106.67 | 1600 | 0.9609 | 0.8410 | 0.8410 |
| 0.023 | 120.0 | 1800 | 0.9775 | 0.8284 | 0.8285 |
| 0.0186 | 133.33 | 2000 | 1.0022 | 0.8285 | 0.8285 |
| 0.0173 | 146.67 | 2200 | 1.1680 | 0.8319 | 0.8326 |
| 0.0146 | 160.0 | 2400 | 1.0761 | 0.8448 | 0.8452 |
| 0.0144 | 173.33 | 2600 | 1.0539 | 0.8326 | 0.8326 |
| 0.0133 | 186.67 | 2800 | 1.0126 | 0.8409 | 0.8410 |
| 0.0117 | 200.0 | 3000 | 1.0532 | 0.8368 | 0.8368 |
| 0.0092 | 213.33 | 3200 | 1.1663 | 0.8284 | 0.8285 |
| 0.0094 | 226.67 | 3400 | 1.1239 | 0.8284 | 0.8285 |
| 0.0095 | 240.0 | 3600 | 1.2638 | 0.8530 | 0.8536 |
| 0.0125 | 253.33 | 3800 | 1.0309 | 0.8368 | 0.8368 |
| 0.0064 | 266.67 | 4000 | 1.1432 | 0.8409 | 0.8410 |
| 0.007 | 280.0 | 4200 | 1.1040 | 0.8326 | 0.8326 |
| 0.0079 | 293.33 | 4400 | 1.1816 | 0.8492 | 0.8494 |
| 0.0067 | 306.67 | 4600 | 1.2173 | 0.8492 | 0.8494 |
| 0.0068 | 320.0 | 4800 | 1.2133 | 0.8534 | 0.8536 |
| 0.0057 | 333.33 | 5000 | 1.2212 | 0.8368 | 0.8368 |
| 0.0046 | 346.67 | 5200 | 1.3225 | 0.8701 | 0.8703 |
| 0.0048 | 360.0 | 5400 | 1.2958 | 0.8410 | 0.8410 |
| 0.0055 | 373.33 | 5600 | 1.2440 | 0.8368 | 0.8368 |
| 0.0059 | 386.67 | 5800 | 1.2122 | 0.8702 | 0.8703 |
| 0.0046 | 400.0 | 6000 | 1.3290 | 0.8700 | 0.8703 |
| 0.0033 | 413.33 | 6200 | 1.2050 | 0.8536 | 0.8536 |
| 0.0046 | 426.67 | 6400 | 1.2307 | 0.8576 | 0.8577 |
| 0.0035 | 440.0 | 6600 | 1.2843 | 0.8614 | 0.8619 |
| 0.0034 | 453.33 | 6800 | 1.2843 | 0.8615 | 0.8619 |
| 0.0031 | 466.67 | 7000 | 1.2826 | 0.8451 | 0.8452 |
| 0.0031 | 480.0 | 7200 | 1.3812 | 0.8368 | 0.8368 |
| 0.0032 | 493.33 | 7400 | 1.2944 | 0.8617 | 0.8619 |
| 0.0025 | 506.67 | 7600 | 1.2717 | 0.8536 | 0.8536 |
| 0.0026 | 520.0 | 7800 | 1.2352 | 0.8575 | 0.8577 |
| 0.0016 | 533.33 | 8000 | 1.3468 | 0.8617 | 0.8619 |
| 0.0027 | 546.67 | 8200 | 1.4685 | 0.8493 | 0.8494 |
| 0.002 | 560.0 | 8400 | 1.4490 | 0.8576 | 0.8577 |
| 0.0024 | 573.33 | 8600 | 1.3509 | 0.8618 | 0.8619 |
| 0.0022 | 586.67 | 8800 | 1.3610 | 0.8659 | 0.8661 |
| 0.0018 | 600.0 | 9000 | 1.4007 | 0.8494 | 0.8494 |
| 0.0019 | 613.33 | 9200 | 1.4635 | 0.8575 | 0.8577 |
| 0.0014 | 626.67 | 9400 | 1.4575 | 0.8577 | 0.8577 |
| 0.0018 | 640.0 | 9600 | 1.4003 | 0.8660 | 0.8661 |
| 0.002 | 653.33 | 9800 | 1.3962 | 0.8619 | 0.8619 |
| 0.0018 | 666.67 | 10000 | 1.3987 | 0.8577 | 0.8577 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_mouse_3-seqsight_16384_512_56M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_mouse_3-seqsight_16384_512_56M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:09:41+00:00 |
text-generation | transformers | <a href="https://www.gradient.ai" target="_blank"><img src="https://cdn-uploads.huggingface.co/production/uploads/655bb613e8a8971e89944f3e/TSa3V8YpoVagnTYgxiLaO.png" width="200"/></a>
# Llama-3 8B Gradient Instruct 1048k
Gradient incorporates your data to deploy autonomous assistants that power critical operations across your business. If you're looking to build custom AI models or agents, email us a message [email protected].
For more info see our [End-to-end development service for custom LLMs and AI systems](https://gradient.ai/development-lab)
This model extends LLama-3 8B's context length from 8k to > 1040K, developed by Gradient, sponsored by compute from [Crusoe Energy](https://huggingface.co/crusoeai). It demonstrates that SOTA LLMs can learn to operate on long context with minimal training by appropriately adjusting RoPE theta. We trained on 830M tokens for this stage, and 1.4B tokens total for all stages, which is < 0.01% of Llama-3's original pre-training data.
**Approach:**
- [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) as the base
- NTK-aware interpolation [1] to initialize an optimal schedule for RoPE theta, followed by empirical RoPE theta optimization
- Progressive training on increasing context lengths, similar to [Large World Model](https://huggingface.co/LargeWorldModel) [2] (See details below)
**Infra:**
We build on top of the EasyContext Blockwise RingAttention library [3] to scalably and efficiently train on contexts up to 1048k tokens on [Crusoe Energy](https://huggingface.co/crusoeai) high performance L40S cluster.
Notably, we layered parallelism on top of Ring Attention with a custom network topology to better leverage large GPU clusters in the face of network bottlenecks from passing many KV blocks between devices. This gave us a 33x speedup in model training (compare 524k and 1048k to 65k and 262k in the table below).
**Data:**
For training data, we generate long contexts by augmenting [SlimPajama](https://huggingface.co/datasets/cerebras/SlimPajama-627B).
**Progressive Training Details:**
| | 65K | 262K | 524k | 1048k |
|------------------------|-----------|-----------|-----------|-----------|
| Initialize From | LLaMA-3 8B| 65K | 262K | 524k |
| Sequence Length 2^N | 16 | 18 | 19 | 20 |
| RoPE theta | 15.3 M | 207.1 M | 1.06B | 2.80B |
| Batch Size | 1 | 1 | 16 | 16 |
| Gradient Accumulation Steps | 32 | 16 | 1 | 1 |
| Steps | 30 | 24 | 50 | 50 |
| Total Tokens | 62914560 | 100663296 | 419430400 | 838860800 |
| Learning Rate | 2.00E-05 | 2.00E-05 | 2.00E-05 | 2.00E-05 |
| # GPUs | 8 | 32 | 512 | 512 |
| GPU Type | NVIDIA L40S | NVIDIA L40S | NVIDIA L40S | NVIDIA L40S |
| Minutes to Train (Wall)| 202 | 555 | 61 | 87 |
**Quants**:
- [GGUF](https://huggingface.co/crusoeai/Llama-3-8B-Instruct-1048k-GGUF)
- [MLX-4bit](https://huggingface.co/mlx-community/Llama-3-8B-Instruct-1048k-4bit)
## The Gradient AI Team
https://gradient.ai/
Gradient is accelerating AI transformation across industries. Our AI Foundry incorporates your data to deploy autonomous assistants that power critical operations across your business.
## Contact Us
Drop an email to [[email protected]](mailto:[email protected])
## References
[1] Peng, Bowen, et al. "Yarn: Efficient context window extension of large language models." arXiv preprint arXiv:2309.00071 (2023).
[2] Liu, Hao, et al. "World Model on Million-Length Video And Language With RingAttention." arXiv preprint arXiv:2402.08268 (2024).
[3] https://github.com/jzhang38/EasyContext
----
# Base Model
## Model Details
Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.
**Model developers** Meta
**Variations** Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.
**Input** Models input text only.
**Output** Models generate text and code only.
**Model Architecture** Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.
<table>
<tr>
<td>
</td>
<td><strong>Training Data</strong>
</td>
<td><strong>Params</strong>
</td>
<td><strong>Context length</strong>
</td>
<td><strong>GQA</strong>
</td>
<td><strong>Token count</strong>
</td>
<td><strong>Knowledge cutoff</strong>
</td>
</tr>
<tr>
<td rowspan="2" >Llama 3
</td>
<td rowspan="2" >A new mix of publicly available online data.
</td>
<td>8B
</td>
<td>8k
</td>
<td>Yes
</td>
<td rowspan="2" >15T+
</td>
<td>March, 2023
</td>
</tr>
<tr>
<td>70B
</td>
<td>8k
</td>
<td>Yes
</td>
<td>December, 2023
</td>
</tr>
</table>
**Llama 3 family of models**. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.
**Model Release Date** April 18, 2024.
**Status** This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.
**License** A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license)
Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes).
## Intended Use
**Intended Use Cases** Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.
**Out-of-scope** Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English**.
**Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.
## How to use
This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original `llama3` codebase.
### Use with transformers
You can run conversational inference using the Transformers pipeline abstraction, or by leveraging the Auto classes with the `generate()` function. Let's see examples of both.
#### Transformers pipeline
```python
import transformers
import torch
model_id = "meta-llama/Meta-Llama-3-8B-Instruct"
pipeline = transformers.pipeline(
"text-generation",
model=model_id,
model_kwargs={"torch_dtype": torch.bfloat16},
device_map="auto",
)
messages = [
{"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
{"role": "user", "content": "Who are you?"},
]
prompt = pipeline.tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
terminators = [
pipeline.tokenizer.eos_token_id,
pipeline.tokenizer.convert_tokens_to_ids("<|eot_id|>")
]
outputs = pipeline(
prompt,
max_new_tokens=256,
eos_token_id=terminators,
do_sample=True,
temperature=0.6,
top_p=0.9,
)
print(outputs[0]["generated_text"][len(prompt):])
```
#### Transformers AutoModelForCausalLM
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
model_id = "meta-llama/Meta-Llama-3-8B-Instruct"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.bfloat16,
device_map="auto",
)
messages = [
{"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
{"role": "user", "content": "Who are you?"},
]
input_ids = tokenizer.apply_chat_template(
messages,
add_generation_prompt=True,
return_tensors="pt"
).to(model.device)
terminators = [
tokenizer.eos_token_id,
tokenizer.convert_tokens_to_ids("<|eot_id|>")
]
outputs = model.generate(
input_ids,
max_new_tokens=256,
eos_token_id=terminators,
do_sample=True,
temperature=0.6,
top_p=0.9,
)
response = outputs[0][input_ids.shape[-1]:]
print(tokenizer.decode(response, skip_special_tokens=True))
```
### Use with `llama3`
Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3)
To download Original checkpoints, see the example command below leveraging `huggingface-cli`:
```
huggingface-cli download meta-llama/Meta-Llama-3-8B-Instruct --include "original/*" --local-dir Meta-Llama-3-8B-Instruct
```
For Hugging Face support, we recommend using transformers or TGI, but a similar command works.
## Hardware and Software
**Training Factors** We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.
**Carbon Footprint Pretraining utilized a cumulative** 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.
<table>
<tr>
<td>
</td>
<td><strong>Time (GPU hours)</strong>
</td>
<td><strong>Power Consumption (W)</strong>
</td>
<td><strong>Carbon Emitted(tCO2eq)</strong>
</td>
</tr>
<tr>
<td>Llama 3 8B
</td>
<td>1.3M
</td>
<td>700
</td>
<td>390
</td>
</tr>
<tr>
<td>Llama 3 70B
</td>
<td>6.4M
</td>
<td>700
</td>
<td>1900
</td>
</tr>
<tr>
<td>Total
</td>
<td>7.7M
</td>
<td>
</td>
<td>2290
</td>
</tr>
</table>
**CO2 emissions during pre-training**. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.
## Training Data
**Overview** Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.
**Data Freshness** The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.
## Benchmarks
In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md).
### Base pretrained models
<table>
<tr>
<td><strong>Category</strong>
</td>
<td><strong>Benchmark</strong>
</td>
<td><strong>Llama 3 8B</strong>
</td>
<td><strong>Llama2 7B</strong>
</td>
<td><strong>Llama2 13B</strong>
</td>
<td><strong>Llama 3 70B</strong>
</td>
<td><strong>Llama2 70B</strong>
</td>
</tr>
<tr>
<td rowspan="6" >General
</td>
<td>MMLU (5-shot)
</td>
<td>66.6
</td>
<td>45.7
</td>
<td>53.8
</td>
<td>79.5
</td>
<td>69.7
</td>
</tr>
<tr>
<td>AGIEval English (3-5 shot)
</td>
<td>45.9
</td>
<td>28.8
</td>
<td>38.7
</td>
<td>63.0
</td>
<td>54.8
</td>
</tr>
<tr>
<td>CommonSenseQA (7-shot)
</td>
<td>72.6
</td>
<td>57.6
</td>
<td>67.6
</td>
<td>83.8
</td>
<td>78.7
</td>
</tr>
<tr>
<td>Winogrande (5-shot)
</td>
<td>76.1
</td>
<td>73.3
</td>
<td>75.4
</td>
<td>83.1
</td>
<td>81.8
</td>
</tr>
<tr>
<td>BIG-Bench Hard (3-shot, CoT)
</td>
<td>61.1
</td>
<td>38.1
</td>
<td>47.0
</td>
<td>81.3
</td>
<td>65.7
</td>
</tr>
<tr>
<td>ARC-Challenge (25-shot)
</td>
<td>78.6
</td>
<td>53.7
</td>
<td>67.6
</td>
<td>93.0
</td>
<td>85.3
</td>
</tr>
<tr>
<td>Knowledge reasoning
</td>
<td>TriviaQA-Wiki (5-shot)
</td>
<td>78.5
</td>
<td>72.1
</td>
<td>79.6
</td>
<td>89.7
</td>
<td>87.5
</td>
</tr>
<tr>
<td rowspan="4" >Reading comprehension
</td>
<td>SQuAD (1-shot)
</td>
<td>76.4
</td>
<td>72.2
</td>
<td>72.1
</td>
<td>85.6
</td>
<td>82.6
</td>
</tr>
<tr>
<td>QuAC (1-shot, F1)
</td>
<td>44.4
</td>
<td>39.6
</td>
<td>44.9
</td>
<td>51.1
</td>
<td>49.4
</td>
</tr>
<tr>
<td>BoolQ (0-shot)
</td>
<td>75.7
</td>
<td>65.5
</td>
<td>66.9
</td>
<td>79.0
</td>
<td>73.1
</td>
</tr>
<tr>
<td>DROP (3-shot, F1)
</td>
<td>58.4
</td>
<td>37.9
</td>
<td>49.8
</td>
<td>79.7
</td>
<td>70.2
</td>
</tr>
</table>
### Instruction tuned models
<table>
<tr>
<td><strong>Benchmark</strong>
</td>
<td><strong>Llama 3 8B</strong>
</td>
<td><strong>Llama 2 7B</strong>
</td>
<td><strong>Llama 2 13B</strong>
</td>
<td><strong>Llama 3 70B</strong>
</td>
<td><strong>Llama 2 70B</strong>
</td>
</tr>
<tr>
<td>MMLU (5-shot)
</td>
<td>68.4
</td>
<td>34.1
</td>
<td>47.8
</td>
<td>82.0
</td>
<td>52.9
</td>
</tr>
<tr>
<td>GPQA (0-shot)
</td>
<td>34.2
</td>
<td>21.7
</td>
<td>22.3
</td>
<td>39.5
</td>
<td>21.0
</td>
</tr>
<tr>
<td>HumanEval (0-shot)
</td>
<td>62.2
</td>
<td>7.9
</td>
<td>14.0
</td>
<td>81.7
</td>
<td>25.6
</td>
</tr>
<tr>
<td>GSM-8K (8-shot, CoT)
</td>
<td>79.6
</td>
<td>25.7
</td>
<td>77.4
</td>
<td>93.0
</td>
<td>57.5
</td>
</tr>
<tr>
<td>MATH (4-shot, CoT)
</td>
<td>30.0
</td>
<td>3.8
</td>
<td>6.7
</td>
<td>50.4
</td>
<td>11.6
</td>
</tr>
</table>
### Responsibility & Safety
We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.
Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.
Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.
As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started.
#### Llama 3-Instruct
As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.
<span style="text-decoration:underline;">Safety</span>
For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.
<span style="text-decoration:underline;">Refusals</span>
In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.
We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.
#### Responsible release
In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.
Misuse
If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/).
#### Critical risks
<span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)
We have conducted a two fold assessment of the safety of the model in this area:
* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.
* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).
### <span style="text-decoration:underline;">Cyber Security </span>
We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval).
### <span style="text-decoration:underline;">Child Safety</span>
Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.
### Community
Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama).
Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community.
## Ethical Considerations and Limitations
The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.
But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.
Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide)
## Citation instructions
@article{llama3modelcard,
title={Llama 3 Model Card},
author={AI@Meta},
year={2024},
url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md}
}
## Contributors
Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos | {"language": ["en"], "license": "llama3", "tags": ["meta", "llama-3"], "pipeline_tag": "text-generation"} | blockblockblock/Llama-3-8B-Instruct-Gradient-1048k-bpw5-exl2 | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"meta",
"llama-3",
"conversational",
"en",
"license:llama3",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"5-bit",
"region:us"
] | null | 2024-04-30T03:09:50+00:00 |
null | peft |
<!-- 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. -->
# GUE_mouse_2-seqsight_16384_512_56M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_mouse_2](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_mouse_2) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3112
- F1 Score: 0.8810
- Accuracy: 0.8811
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:------:|:-----:|:---------------:|:--------:|:--------:|
| 0.3736 | 9.52 | 200 | 0.3294 | 0.8382 | 0.8384 |
| 0.2764 | 19.05 | 400 | 0.3029 | 0.8472 | 0.8476 |
| 0.2484 | 28.57 | 600 | 0.2972 | 0.8687 | 0.8689 |
| 0.224 | 38.1 | 800 | 0.3012 | 0.8597 | 0.8598 |
| 0.2055 | 47.62 | 1000 | 0.2956 | 0.8689 | 0.8689 |
| 0.1896 | 57.14 | 1200 | 0.3117 | 0.8653 | 0.8659 |
| 0.1743 | 66.67 | 1400 | 0.3398 | 0.8589 | 0.8598 |
| 0.1632 | 76.19 | 1600 | 0.3155 | 0.8626 | 0.8628 |
| 0.1478 | 85.71 | 1800 | 0.3322 | 0.8748 | 0.875 |
| 0.1435 | 95.24 | 2000 | 0.3429 | 0.8533 | 0.8537 |
| 0.1314 | 104.76 | 2200 | 0.3569 | 0.8656 | 0.8659 |
| 0.1202 | 114.29 | 2400 | 0.3536 | 0.8688 | 0.8689 |
| 0.1133 | 123.81 | 2600 | 0.3958 | 0.8532 | 0.8537 |
| 0.1084 | 133.33 | 2800 | 0.4056 | 0.8683 | 0.8689 |
| 0.1054 | 142.86 | 3000 | 0.4366 | 0.8651 | 0.8659 |
| 0.0953 | 152.38 | 3200 | 0.4359 | 0.8559 | 0.8567 |
| 0.0931 | 161.9 | 3400 | 0.4335 | 0.8624 | 0.8628 |
| 0.0879 | 171.43 | 3600 | 0.4117 | 0.8687 | 0.8689 |
| 0.0812 | 180.95 | 3800 | 0.4762 | 0.8499 | 0.8506 |
| 0.0763 | 190.48 | 4000 | 0.5332 | 0.8557 | 0.8567 |
| 0.0776 | 200.0 | 4200 | 0.5967 | 0.8432 | 0.8445 |
| 0.0715 | 209.52 | 4400 | 0.5072 | 0.8562 | 0.8567 |
| 0.0704 | 219.05 | 4600 | 0.4737 | 0.8656 | 0.8659 |
| 0.0719 | 228.57 | 4800 | 0.4756 | 0.8593 | 0.8598 |
| 0.0674 | 238.1 | 5000 | 0.5231 | 0.8437 | 0.8445 |
| 0.0622 | 247.62 | 5200 | 0.4919 | 0.8564 | 0.8567 |
| 0.0607 | 257.14 | 5400 | 0.5254 | 0.8501 | 0.8506 |
| 0.0583 | 266.67 | 5600 | 0.5381 | 0.8498 | 0.8506 |
| 0.0586 | 276.19 | 5800 | 0.5637 | 0.8499 | 0.8506 |
| 0.0541 | 285.71 | 6000 | 0.5516 | 0.8468 | 0.8476 |
| 0.0559 | 295.24 | 6200 | 0.5670 | 0.8562 | 0.8567 |
| 0.0537 | 304.76 | 6400 | 0.5428 | 0.8562 | 0.8567 |
| 0.0551 | 314.29 | 6600 | 0.5029 | 0.8656 | 0.8659 |
| 0.0489 | 323.81 | 6800 | 0.5435 | 0.8623 | 0.8628 |
| 0.0495 | 333.33 | 7000 | 0.5511 | 0.8562 | 0.8567 |
| 0.0504 | 342.86 | 7200 | 0.5463 | 0.8593 | 0.8598 |
| 0.0485 | 352.38 | 7400 | 0.5462 | 0.8654 | 0.8659 |
| 0.0438 | 361.9 | 7600 | 0.5615 | 0.8623 | 0.8628 |
| 0.0478 | 371.43 | 7800 | 0.5537 | 0.8563 | 0.8567 |
| 0.0454 | 380.95 | 8000 | 0.5433 | 0.8593 | 0.8598 |
| 0.0465 | 390.48 | 8200 | 0.5488 | 0.8563 | 0.8567 |
| 0.0438 | 400.0 | 8400 | 0.5694 | 0.8592 | 0.8598 |
| 0.0404 | 409.52 | 8600 | 0.5840 | 0.8561 | 0.8567 |
| 0.0419 | 419.05 | 8800 | 0.5785 | 0.8623 | 0.8628 |
| 0.0431 | 428.57 | 9000 | 0.6124 | 0.8561 | 0.8567 |
| 0.0416 | 438.1 | 9200 | 0.5836 | 0.8561 | 0.8567 |
| 0.0441 | 447.62 | 9400 | 0.5770 | 0.8561 | 0.8567 |
| 0.0387 | 457.14 | 9600 | 0.5626 | 0.8562 | 0.8567 |
| 0.0406 | 466.67 | 9800 | 0.5803 | 0.8592 | 0.8598 |
| 0.0412 | 476.19 | 10000 | 0.5708 | 0.8592 | 0.8598 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_mouse_2-seqsight_16384_512_56M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_mouse_2-seqsight_16384_512_56M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:10:15+00:00 |
sentence-similarity | transformers |
# LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders
> LLM2Vec is a simple recipe to convert decoder-only LLMs into text encoders. It consists of 3 simple steps: 1) enabling bidirectional attention, 2) masked next token prediction, and 3) unsupervised contrastive learning. The model can be further fine-tuned to achieve state-of-the-art performance.
- **Repository:** https://github.com/McGill-NLP/llm2vec
- **Paper:** https://arxiv.org/abs/2404.05961
## Installation
```bash
pip install llm2vec
```
## Usage
```python
from llm2vec import LLM2Vec
import torch
from transformers import AutoTokenizer, AutoModel, AutoConfig
from peft import PeftModel
# Loading base Mistral model, along with custom code that enables bidirectional connections in decoder-only LLMs.
tokenizer = AutoTokenizer.from_pretrained(
"McGill-NLP/LLM2Vec-Meta-Llama-3-8B-Instruct-mntp"
)
config = AutoConfig.from_pretrained(
"McGill-NLP/LLM2Vec-Meta-Llama-3-8B-Instruct-mntp", trust_remote_code=True
)
model = AutoModel.from_pretrained(
"McGill-NLP/LLM2Vec-Meta-Llama-3-8B-Instruct-mntp",
trust_remote_code=True,
config=config,
torch_dtype=torch.bfloat16,
device_map="cuda" if torch.cuda.is_available() else "cpu",
)
# Loading MNTP (Masked Next Token Prediction) model.
model = PeftModel.from_pretrained(
model,
"McGill-NLP/LLM2Vec-Meta-Llama-3-8B-Instruct-mntp",
)
# Wrapper for encoding and pooling operations
l2v = LLM2Vec(model, tokenizer, pooling_mode="mean", max_length=512)
# Encoding queries using instructions
instruction = (
"Given a web search query, retrieve relevant passages that answer the query:"
)
queries = [
[instruction, "how much protein should a female eat"],
[instruction, "summit define"],
]
q_reps = l2v.encode(queries)
# Encoding documents. Instruction are not required for documents
documents = [
"As a general guideline, the CDC's average requirement of protein for women ages 19 to 70 is 46 grams per day. But, as you can see from this chart, you'll need to increase that if you're expecting or training for a marathon. Check out the chart below to see how much protein you should be eating each day.",
"Definition of summit for English Language Learners. : 1 the highest point of a mountain : the top of a mountain. : 2 the highest level. : 3 a meeting or series of meetings between the leaders of two or more governments.",
]
d_reps = l2v.encode(documents)
# Compute cosine similarity
q_reps_norm = torch.nn.functional.normalize(q_reps, p=2, dim=1)
d_reps_norm = torch.nn.functional.normalize(d_reps, p=2, dim=1)
cos_sim = torch.mm(q_reps_norm, d_reps_norm.transpose(0, 1))
print(cos_sim)
"""
tensor([[0.7740, 0.5580],
[0.4845, 0.4993]])
"""
```
## Questions
If you have any question about the code, feel free to email Parishad (`[email protected]`) and Vaibhav (`[email protected]`). | {"language": ["en"], "license": "mit", "library_name": "transformers", "tags": ["text-embedding", "embeddings", "information-retrieval", "beir", "text-classification", "language-model", "text-clustering", "text-semantic-similarity", "text-evaluation", "text-reranking", "feature-extraction", "sentence-similarity", "Sentence Similarity", "natural_questions", "ms_marco", "fever", "hotpot_qa", "mteb"], "pipeline_tag": "sentence-similarity"} | McGill-NLP/LLM2Vec-Meta-Llama-3-8B-Instruct-mntp | null | [
"transformers",
"safetensors",
"llama",
"feature-extraction",
"text-embedding",
"embeddings",
"information-retrieval",
"beir",
"text-classification",
"language-model",
"text-clustering",
"text-semantic-similarity",
"text-evaluation",
"text-reranking",
"sentence-similarity",
"Sentence Similarity",
"natural_questions",
"ms_marco",
"fever",
"hotpot_qa",
"mteb",
"custom_code",
"en",
"arxiv:2404.05961",
"license:mit",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:10:31+00:00 |
null | peft |
<!-- 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. -->
# GUE_mouse_2-seqsight_16384_512_56M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_mouse_2](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_mouse_2) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6632
- F1 Score: 0.8841
- Accuracy: 0.8841
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:------:|:-----:|:---------------:|:--------:|:--------:|
| 0.3425 | 9.52 | 200 | 0.3026 | 0.8567 | 0.8567 |
| 0.2324 | 19.05 | 400 | 0.2745 | 0.8659 | 0.8659 |
| 0.1826 | 28.57 | 600 | 0.3028 | 0.8746 | 0.875 |
| 0.1478 | 38.1 | 800 | 0.3116 | 0.8902 | 0.8902 |
| 0.117 | 47.62 | 1000 | 0.3371 | 0.9085 | 0.9085 |
| 0.0884 | 57.14 | 1200 | 0.4229 | 0.8868 | 0.8872 |
| 0.079 | 66.67 | 1400 | 0.3760 | 0.9115 | 0.9116 |
| 0.0666 | 76.19 | 1600 | 0.4414 | 0.8869 | 0.8872 |
| 0.0559 | 85.71 | 1800 | 0.4145 | 0.9115 | 0.9116 |
| 0.0504 | 95.24 | 2000 | 0.5103 | 0.8775 | 0.8780 |
| 0.0467 | 104.76 | 2200 | 0.4754 | 0.8870 | 0.8872 |
| 0.0396 | 114.29 | 2400 | 0.4569 | 0.8961 | 0.8963 |
| 0.0353 | 123.81 | 2600 | 0.4563 | 0.9116 | 0.9116 |
| 0.0309 | 133.33 | 2800 | 0.5872 | 0.8773 | 0.8780 |
| 0.0312 | 142.86 | 3000 | 0.5229 | 0.8993 | 0.8994 |
| 0.0255 | 152.38 | 3200 | 0.5115 | 0.9023 | 0.9024 |
| 0.029 | 161.9 | 3400 | 0.4790 | 0.9055 | 0.9055 |
| 0.0241 | 171.43 | 3600 | 0.5294 | 0.9024 | 0.9024 |
| 0.0237 | 180.95 | 3800 | 0.5058 | 0.8962 | 0.8963 |
| 0.0175 | 190.48 | 4000 | 0.5896 | 0.8901 | 0.8902 |
| 0.019 | 200.0 | 4200 | 0.7127 | 0.8772 | 0.8780 |
| 0.0171 | 209.52 | 4400 | 0.5494 | 0.8901 | 0.8902 |
| 0.0172 | 219.05 | 4600 | 0.5188 | 0.9116 | 0.9116 |
| 0.0168 | 228.57 | 4800 | 0.5622 | 0.9085 | 0.9085 |
| 0.0165 | 238.1 | 5000 | 0.5741 | 0.8870 | 0.8872 |
| 0.0154 | 247.62 | 5200 | 0.6399 | 0.8932 | 0.8933 |
| 0.0152 | 257.14 | 5400 | 0.5727 | 0.9116 | 0.9116 |
| 0.0135 | 266.67 | 5600 | 0.5515 | 0.8994 | 0.8994 |
| 0.0123 | 276.19 | 5800 | 0.5853 | 0.8902 | 0.8902 |
| 0.0121 | 285.71 | 6000 | 0.5656 | 0.8993 | 0.8994 |
| 0.0116 | 295.24 | 6200 | 0.5896 | 0.9024 | 0.9024 |
| 0.0146 | 304.76 | 6400 | 0.5939 | 0.8993 | 0.8994 |
| 0.0126 | 314.29 | 6600 | 0.5729 | 0.9054 | 0.9055 |
| 0.0088 | 323.81 | 6800 | 0.6025 | 0.8993 | 0.8994 |
| 0.0086 | 333.33 | 7000 | 0.5996 | 0.8994 | 0.8994 |
| 0.0104 | 342.86 | 7200 | 0.6030 | 0.9024 | 0.9024 |
| 0.0092 | 352.38 | 7400 | 0.5908 | 0.9024 | 0.9024 |
| 0.0069 | 361.9 | 7600 | 0.6540 | 0.9024 | 0.9024 |
| 0.01 | 371.43 | 7800 | 0.6252 | 0.8932 | 0.8933 |
| 0.0086 | 380.95 | 8000 | 0.6711 | 0.8932 | 0.8933 |
| 0.009 | 390.48 | 8200 | 0.6450 | 0.8963 | 0.8963 |
| 0.0075 | 400.0 | 8400 | 0.6476 | 0.8963 | 0.8963 |
| 0.0074 | 409.52 | 8600 | 0.6735 | 0.8932 | 0.8933 |
| 0.0072 | 419.05 | 8800 | 0.6793 | 0.8994 | 0.8994 |
| 0.0073 | 428.57 | 9000 | 0.6979 | 0.8869 | 0.8872 |
| 0.0082 | 438.1 | 9200 | 0.6626 | 0.8963 | 0.8963 |
| 0.0081 | 447.62 | 9400 | 0.6683 | 0.8932 | 0.8933 |
| 0.0075 | 457.14 | 9600 | 0.6605 | 0.8933 | 0.8933 |
| 0.0065 | 466.67 | 9800 | 0.6637 | 0.8994 | 0.8994 |
| 0.0076 | 476.19 | 10000 | 0.6640 | 0.8963 | 0.8963 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_mouse_2-seqsight_16384_512_56M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_mouse_2-seqsight_16384_512_56M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:10:31+00:00 |
text-classification | transformers | ## TextAttack Model Card
This `bert` model was fine-tuned using TextAttack. The model was fine-tuned
for 3 epochs with a batch size of 8,
a maximum sequence length of 512, and an initial learning rate of 3e-05.
Since this was a classification task, the model was trained with a cross-entropy loss function.
The best score the model achieved on this task was 0.9556666666666667, as measured by the
eval set accuracy, found after 3 epochs.
For more information, check out [TextAttack on Github](https://github.com/QData/TextAttack). | {"language": ["zh"], "metrics": ["accuracy"], "pipeline_tag": "text-classification"} | WangA/roberta-base-finetuned-jd | null | [
"transformers",
"safetensors",
"bert",
"text-classification",
"zh",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:10:48+00:00 |
null | null | {} | luciusy/ts_qa_dummy_pp2 | null | [
"region:us"
] | null | 2024-04-30T03:10:51+00:00 |
|
text-generation | transformers |
<!-- 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. -->
# ppo_zephyr9
This model is a fine-tuned version of [HuggingFaceH4/mistral-7b-sft-beta](https://huggingface.co/HuggingFaceH4/mistral-7b-sft-beta) on an unknown 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: 3e-06
- train_batch_size: 1
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 32
- total_train_batch_size: 256
- total_eval_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Framework versions
- Transformers 4.40.1
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.19.1
| {"license": "mit", "tags": ["generated_from_trainer"], "base_model": "HuggingFaceH4/mistral-7b-sft-beta", "model-index": [{"name": "ppo_zephyr9", "results": []}]} | vwxyzjn/ppo_zephyr9 | null | [
"transformers",
"tensorboard",
"safetensors",
"mistral",
"text-generation",
"generated_from_trainer",
"conversational",
"base_model:HuggingFaceH4/mistral-7b-sft-beta",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:10:55+00:00 |
null | peft |
<!-- 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. -->
# GUE_mouse_2-seqsight_16384_512_56M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_mouse_2](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_mouse_2) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9767
- F1 Score: 0.8779
- Accuracy: 0.8780
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:------:|:-----:|:---------------:|:--------:|:--------:|
| 0.3144 | 9.52 | 200 | 0.2767 | 0.8963 | 0.8963 |
| 0.1896 | 19.05 | 400 | 0.2933 | 0.8841 | 0.8841 |
| 0.1174 | 28.57 | 600 | 0.3202 | 0.8993 | 0.8994 |
| 0.0783 | 38.1 | 800 | 0.3931 | 0.9085 | 0.9085 |
| 0.0528 | 47.62 | 1000 | 0.4485 | 0.8961 | 0.8963 |
| 0.0346 | 57.14 | 1200 | 0.5750 | 0.8869 | 0.8872 |
| 0.0328 | 66.67 | 1400 | 0.5366 | 0.8870 | 0.8872 |
| 0.0219 | 76.19 | 1600 | 0.7088 | 0.8745 | 0.875 |
| 0.0202 | 85.71 | 1800 | 0.6053 | 0.8993 | 0.8994 |
| 0.0215 | 95.24 | 2000 | 0.5883 | 0.8963 | 0.8963 |
| 0.016 | 104.76 | 2200 | 0.6117 | 0.8871 | 0.8872 |
| 0.0135 | 114.29 | 2400 | 0.6432 | 0.8869 | 0.8872 |
| 0.0097 | 123.81 | 2600 | 0.5925 | 0.9055 | 0.9055 |
| 0.0119 | 133.33 | 2800 | 0.6381 | 0.8933 | 0.8933 |
| 0.0115 | 142.86 | 3000 | 0.5637 | 0.9024 | 0.9024 |
| 0.0067 | 152.38 | 3200 | 0.6507 | 0.8931 | 0.8933 |
| 0.0095 | 161.9 | 3400 | 0.6016 | 0.8933 | 0.8933 |
| 0.0051 | 171.43 | 3600 | 0.6518 | 0.8963 | 0.8963 |
| 0.0084 | 180.95 | 3800 | 0.6275 | 0.8962 | 0.8963 |
| 0.0059 | 190.48 | 4000 | 0.6769 | 0.8994 | 0.8994 |
| 0.0063 | 200.0 | 4200 | 0.6820 | 0.8994 | 0.8994 |
| 0.0056 | 209.52 | 4400 | 0.7624 | 0.8901 | 0.8902 |
| 0.0049 | 219.05 | 4600 | 0.7319 | 0.9024 | 0.9024 |
| 0.0058 | 228.57 | 4800 | 0.7014 | 0.8932 | 0.8933 |
| 0.0049 | 238.1 | 5000 | 0.6703 | 0.8902 | 0.8902 |
| 0.0052 | 247.62 | 5200 | 0.6760 | 0.8901 | 0.8902 |
| 0.0031 | 257.14 | 5400 | 0.6890 | 0.8963 | 0.8963 |
| 0.003 | 266.67 | 5600 | 0.7032 | 0.8933 | 0.8933 |
| 0.0034 | 276.19 | 5800 | 0.7176 | 0.8963 | 0.8963 |
| 0.0031 | 285.71 | 6000 | 0.8314 | 0.8870 | 0.8872 |
| 0.0042 | 295.24 | 6200 | 0.8368 | 0.8839 | 0.8841 |
| 0.0035 | 304.76 | 6400 | 0.7048 | 0.8994 | 0.8994 |
| 0.0027 | 314.29 | 6600 | 0.7305 | 0.8994 | 0.8994 |
| 0.0016 | 323.81 | 6800 | 0.7220 | 0.8902 | 0.8902 |
| 0.0031 | 333.33 | 7000 | 0.7453 | 0.9055 | 0.9055 |
| 0.0022 | 342.86 | 7200 | 0.8379 | 0.8806 | 0.8811 |
| 0.0027 | 352.38 | 7400 | 0.7857 | 0.8993 | 0.8994 |
| 0.0015 | 361.9 | 7600 | 0.8246 | 0.8900 | 0.8902 |
| 0.0028 | 371.43 | 7800 | 0.6804 | 0.8872 | 0.8872 |
| 0.0021 | 380.95 | 8000 | 0.7188 | 0.9024 | 0.9024 |
| 0.0018 | 390.48 | 8200 | 0.7816 | 0.9085 | 0.9085 |
| 0.0014 | 400.0 | 8400 | 0.7669 | 0.9054 | 0.9055 |
| 0.0011 | 409.52 | 8600 | 0.7935 | 0.9055 | 0.9055 |
| 0.0012 | 419.05 | 8800 | 0.7798 | 0.9115 | 0.9116 |
| 0.0014 | 428.57 | 9000 | 0.7551 | 0.9085 | 0.9085 |
| 0.001 | 438.1 | 9200 | 0.8028 | 0.9024 | 0.9024 |
| 0.001 | 447.62 | 9400 | 0.7968 | 0.9085 | 0.9085 |
| 0.0016 | 457.14 | 9600 | 0.7802 | 0.8933 | 0.8933 |
| 0.0015 | 466.67 | 9800 | 0.7738 | 0.8994 | 0.8994 |
| 0.0009 | 476.19 | 10000 | 0.7851 | 0.8963 | 0.8963 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_mouse_2-seqsight_16384_512_56M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_mouse_2-seqsight_16384_512_56M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:11:03+00:00 |
null | peft |
<!-- 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. -->
# GUE_splice_reconstructed-seqsight_16384_512_56M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_splice_reconstructed](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_splice_reconstructed) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3818
- F1 Score: 0.8522
- Accuracy: 0.8512
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.9326 | 0.7 | 200 | 0.8935 | 0.5689 | 0.5673 |
| 0.7686 | 1.4 | 400 | 0.6592 | 0.7110 | 0.7098 |
| 0.5517 | 2.1 | 600 | 0.5339 | 0.7707 | 0.7696 |
| 0.4981 | 2.8 | 800 | 0.5505 | 0.7653 | 0.7652 |
| 0.4813 | 3.5 | 1000 | 0.4881 | 0.7906 | 0.7896 |
| 0.4611 | 4.2 | 1200 | 0.5080 | 0.7784 | 0.7773 |
| 0.4497 | 4.9 | 1400 | 0.4766 | 0.7984 | 0.7975 |
| 0.4436 | 5.59 | 1600 | 0.4709 | 0.8058 | 0.8049 |
| 0.4357 | 6.29 | 1800 | 0.5207 | 0.7741 | 0.7740 |
| 0.4275 | 6.99 | 2000 | 0.4666 | 0.8038 | 0.8025 |
| 0.4179 | 7.69 | 2200 | 0.4341 | 0.8213 | 0.8203 |
| 0.418 | 8.39 | 2400 | 0.4692 | 0.8017 | 0.8010 |
| 0.4147 | 9.09 | 2600 | 0.4781 | 0.8023 | 0.8007 |
| 0.4109 | 9.79 | 2800 | 0.4568 | 0.8096 | 0.8086 |
| 0.4083 | 10.49 | 3000 | 0.4375 | 0.8217 | 0.8207 |
| 0.3948 | 11.19 | 3200 | 0.4755 | 0.8069 | 0.8056 |
| 0.3983 | 11.89 | 3400 | 0.4439 | 0.8173 | 0.8161 |
| 0.3922 | 12.59 | 3600 | 0.4239 | 0.8314 | 0.8306 |
| 0.392 | 13.29 | 3800 | 0.4360 | 0.8211 | 0.8198 |
| 0.3875 | 13.99 | 4000 | 0.4609 | 0.8142 | 0.8130 |
| 0.3787 | 14.69 | 4200 | 0.4475 | 0.8174 | 0.8161 |
| 0.3832 | 15.38 | 4400 | 0.4102 | 0.8357 | 0.8347 |
| 0.3836 | 16.08 | 4600 | 0.4751 | 0.8052 | 0.8038 |
| 0.3793 | 16.78 | 4800 | 0.4253 | 0.8304 | 0.8292 |
| 0.3674 | 17.48 | 5000 | 0.4559 | 0.8172 | 0.8159 |
| 0.3771 | 18.18 | 5200 | 0.4016 | 0.8438 | 0.8426 |
| 0.3711 | 18.88 | 5400 | 0.4195 | 0.8341 | 0.8330 |
| 0.3668 | 19.58 | 5600 | 0.3937 | 0.8459 | 0.8450 |
| 0.3749 | 20.28 | 5800 | 0.4048 | 0.8425 | 0.8415 |
| 0.366 | 20.98 | 6000 | 0.3878 | 0.8483 | 0.8474 |
| 0.3679 | 21.68 | 6200 | 0.4200 | 0.8349 | 0.8338 |
| 0.366 | 22.38 | 6400 | 0.4078 | 0.8361 | 0.8349 |
| 0.3588 | 23.08 | 6600 | 0.4112 | 0.8359 | 0.8347 |
| 0.362 | 23.78 | 6800 | 0.4112 | 0.8386 | 0.8376 |
| 0.3618 | 24.48 | 7000 | 0.4182 | 0.8350 | 0.8338 |
| 0.3621 | 25.17 | 7200 | 0.4097 | 0.8384 | 0.8371 |
| 0.3607 | 25.87 | 7400 | 0.4061 | 0.8387 | 0.8376 |
| 0.3557 | 26.57 | 7600 | 0.4164 | 0.8336 | 0.8323 |
| 0.3558 | 27.27 | 7800 | 0.4124 | 0.8389 | 0.8378 |
| 0.3539 | 27.97 | 8000 | 0.4111 | 0.8385 | 0.8374 |
| 0.3592 | 28.67 | 8200 | 0.4074 | 0.8408 | 0.8398 |
| 0.3483 | 29.37 | 8400 | 0.4036 | 0.8430 | 0.8417 |
| 0.3477 | 30.07 | 8600 | 0.3888 | 0.8500 | 0.8490 |
| 0.3514 | 30.77 | 8800 | 0.4113 | 0.8401 | 0.8389 |
| 0.3496 | 31.47 | 9000 | 0.4043 | 0.8425 | 0.8413 |
| 0.3503 | 32.17 | 9200 | 0.4026 | 0.8432 | 0.8420 |
| 0.3484 | 32.87 | 9400 | 0.3970 | 0.8435 | 0.8424 |
| 0.3508 | 33.57 | 9600 | 0.4017 | 0.8422 | 0.8411 |
| 0.3498 | 34.27 | 9800 | 0.3975 | 0.8438 | 0.8426 |
| 0.3513 | 34.97 | 10000 | 0.3993 | 0.8435 | 0.8424 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_splice_reconstructed-seqsight_16384_512_56M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_splice_reconstructed-seqsight_16384_512_56M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:11:11+00:00 |
text2text-generation | transformers | {} | paulh27/xsum_aligned_smallT5_cont1 | null | [
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:12:10+00:00 |
|
null | null | {} | BaironGx/Asistente | null | [
"region:us"
] | null | 2024-04-30T03:12:18+00:00 |
|
null | transformers |
# Uploaded model
- **Developed by:** 1024m
- **License:** apache-2.0
- **Finetuned from model :** unsloth/llama-3-8b-bnb-4bit
This llama model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
| {"language": ["en"], "license": "apache-2.0", "tags": ["text-generation-inference", "transformers", "unsloth", "llama", "trl"], "base_model": "unsloth/llama-3-8b-bnb-4bit"} | 1024m/LLAMA3-SMM4H-Task6-LoRA | null | [
"transformers",
"safetensors",
"text-generation-inference",
"unsloth",
"llama",
"trl",
"en",
"base_model:unsloth/llama-3-8b-bnb-4bit",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:12:49+00:00 |
null | peft |
<!-- 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. -->
# out
This model is a fine-tuned version of [google/gemma-1.1-2b-it](https://huggingface.co/google/gemma-1.1-2b-it) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 5.1992
## 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: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.03
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| No log | 1.0 | 1 | 5.1991 |
| No log | 2.0 | 3 | 5.1992 |
### Framework versions
- PEFT 0.10.0
- Transformers 4.40.0
- Pytorch 2.2.1+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1 | {"license": "gemma", "library_name": "peft", "tags": ["trl", "sft", "generated_from_trainer"], "base_model": "google/gemma-1.1-2b-it", "model-index": [{"name": "out", "results": []}]} | cohesionet/out | null | [
"peft",
"tensorboard",
"safetensors",
"trl",
"sft",
"generated_from_trainer",
"base_model:google/gemma-1.1-2b-it",
"license:gemma",
"region:us"
] | null | 2024-04-30T03:14:16+00:00 |
text-generation | transformers |
# Uploaded model
- **Developed by:** 1024m
- **License:** apache-2.0
- **Finetuned from model :** unsloth/llama-3-8b-bnb-4bit
This llama model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
| {"language": ["en"], "license": "apache-2.0", "tags": ["text-generation-inference", "transformers", "unsloth", "llama", "trl"], "base_model": "unsloth/llama-3-8b-bnb-4bit"} | 1024m/LLAMA3-SMM4H-Task6-4bit | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"text-generation-inference",
"unsloth",
"trl",
"en",
"base_model:unsloth/llama-3-8b-bnb-4bit",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"8-bit",
"region:us"
] | null | 2024-04-30T03:14:19+00:00 |
text-generation | transformers |
<!-- 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. -->
# llama2-7b-dpo-full-sft-wo-healthsearch_qa
This model is a fine-tuned version of [Minbyul/llama2-7b-wo-healthsearch_qa-sft](https://huggingface.co/Minbyul/llama2-7b-wo-healthsearch_qa-sft) on the HuggingFaceH4/ultrafeedback_binarized dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6914
- Rewards/chosen: -0.0016
- Rewards/rejected: -0.0049
- Rewards/accuracies: 0.6105
- Rewards/margins: 0.0033
- Logps/rejected: -671.7160
- Logps/chosen: -334.7261
- Logits/rejected: -0.1596
- Logits/chosen: -0.4926
## 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-07
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1
### Training results
### Framework versions
- Transformers 4.39.0.dev0
- Pytorch 2.1.2
- Datasets 2.14.6
- Tokenizers 0.15.2
| {"tags": ["alignment-handbook", "trl", "dpo", "generated_from_trainer", "trl", "dpo", "generated_from_trainer"], "datasets": ["HuggingFaceH4/ultrafeedback_binarized"], "base_model": "Minbyul/llama2-7b-wo-healthsearch_qa-sft", "model-index": [{"name": "llama2-7b-dpo-full-sft-wo-healthsearch_qa", "results": []}]} | Minbyul/llama2-7b-dpo-full-sft-wo-healthsearch_qa | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"alignment-handbook",
"trl",
"dpo",
"generated_from_trainer",
"dataset:HuggingFaceH4/ultrafeedback_binarized",
"base_model:Minbyul/llama2-7b-wo-healthsearch_qa-sft",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:16:12+00:00 |
text2text-generation | transformers | {} | paulh27/cnn_aligned_smallT5_cont1 | null | [
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:17:43+00:00 |
|
null | null | {} | iqsq/planetn6 | null | [
"region:us"
] | null | 2024-04-30T03:19:04+00:00 |
|
null | peft |
<!-- 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. -->
# GUE_splice_reconstructed-seqsight_16384_512_56M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_splice_reconstructed](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_splice_reconstructed) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3406
- F1 Score: 0.8729
- Accuracy: 0.8722
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.8907 | 0.7 | 200 | 0.6928 | 0.6954 | 0.6927 |
| 0.5207 | 1.4 | 400 | 0.5017 | 0.7865 | 0.7854 |
| 0.4602 | 2.1 | 600 | 0.4686 | 0.8023 | 0.8014 |
| 0.4252 | 2.8 | 800 | 0.4887 | 0.7939 | 0.7935 |
| 0.4127 | 3.5 | 1000 | 0.3990 | 0.8398 | 0.8391 |
| 0.3896 | 4.2 | 1200 | 0.4297 | 0.8224 | 0.8214 |
| 0.3825 | 4.9 | 1400 | 0.3971 | 0.8392 | 0.8382 |
| 0.3724 | 5.59 | 1600 | 0.4076 | 0.8370 | 0.8365 |
| 0.3604 | 6.29 | 1800 | 0.4270 | 0.8275 | 0.8266 |
| 0.3545 | 6.99 | 2000 | 0.3876 | 0.8489 | 0.8481 |
| 0.3478 | 7.69 | 2200 | 0.4005 | 0.8417 | 0.8409 |
| 0.3434 | 8.39 | 2400 | 0.3949 | 0.8442 | 0.8435 |
| 0.3366 | 9.09 | 2600 | 0.4114 | 0.8403 | 0.8391 |
| 0.3316 | 9.79 | 2800 | 0.3838 | 0.8511 | 0.8503 |
| 0.3328 | 10.49 | 3000 | 0.3818 | 0.8519 | 0.8512 |
| 0.3195 | 11.19 | 3200 | 0.4075 | 0.8423 | 0.8415 |
| 0.3197 | 11.89 | 3400 | 0.3874 | 0.8487 | 0.8479 |
| 0.3127 | 12.59 | 3600 | 0.3800 | 0.8499 | 0.8492 |
| 0.3134 | 13.29 | 3800 | 0.3666 | 0.8593 | 0.8584 |
| 0.3089 | 13.99 | 4000 | 0.3947 | 0.8491 | 0.8483 |
| 0.2997 | 14.69 | 4200 | 0.3748 | 0.8543 | 0.8536 |
| 0.3017 | 15.38 | 4400 | 0.3666 | 0.8595 | 0.8588 |
| 0.3025 | 16.08 | 4600 | 0.4096 | 0.8437 | 0.8428 |
| 0.2982 | 16.78 | 4800 | 0.3700 | 0.8573 | 0.8564 |
| 0.2839 | 17.48 | 5000 | 0.3930 | 0.8523 | 0.8514 |
| 0.2912 | 18.18 | 5200 | 0.3600 | 0.8637 | 0.8630 |
| 0.2847 | 18.88 | 5400 | 0.3670 | 0.8650 | 0.8643 |
| 0.2806 | 19.58 | 5600 | 0.3562 | 0.8654 | 0.8648 |
| 0.2903 | 20.28 | 5800 | 0.3621 | 0.8653 | 0.8645 |
| 0.2783 | 20.98 | 6000 | 0.3471 | 0.8700 | 0.8694 |
| 0.2785 | 21.68 | 6200 | 0.3739 | 0.8587 | 0.8580 |
| 0.2766 | 22.38 | 6400 | 0.3602 | 0.8619 | 0.8610 |
| 0.2703 | 23.08 | 6600 | 0.3539 | 0.8698 | 0.8691 |
| 0.2729 | 23.78 | 6800 | 0.3677 | 0.8624 | 0.8617 |
| 0.2694 | 24.48 | 7000 | 0.3715 | 0.8619 | 0.8612 |
| 0.2678 | 25.17 | 7200 | 0.3587 | 0.8651 | 0.8643 |
| 0.2699 | 25.87 | 7400 | 0.3544 | 0.8670 | 0.8663 |
| 0.2673 | 26.57 | 7600 | 0.3604 | 0.8630 | 0.8621 |
| 0.2648 | 27.27 | 7800 | 0.3576 | 0.8670 | 0.8663 |
| 0.2625 | 27.97 | 8000 | 0.3624 | 0.8627 | 0.8619 |
| 0.266 | 28.67 | 8200 | 0.3541 | 0.8677 | 0.8669 |
| 0.259 | 29.37 | 8400 | 0.3555 | 0.8675 | 0.8667 |
| 0.2572 | 30.07 | 8600 | 0.3500 | 0.8705 | 0.8698 |
| 0.2599 | 30.77 | 8800 | 0.3600 | 0.8647 | 0.8639 |
| 0.2576 | 31.47 | 9000 | 0.3600 | 0.8656 | 0.8648 |
| 0.2594 | 32.17 | 9200 | 0.3531 | 0.8685 | 0.8678 |
| 0.2546 | 32.87 | 9400 | 0.3637 | 0.8650 | 0.8643 |
| 0.2543 | 33.57 | 9600 | 0.3679 | 0.8622 | 0.8615 |
| 0.2581 | 34.27 | 9800 | 0.3558 | 0.8677 | 0.8669 |
| 0.2509 | 34.97 | 10000 | 0.3580 | 0.8661 | 0.8654 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_splice_reconstructed-seqsight_16384_512_56M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_splice_reconstructed-seqsight_16384_512_56M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:19:14+00:00 |
null | peft |
<!-- 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. -->
# GUE_splice_reconstructed-seqsight_16384_512_56M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_splice_reconstructed](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_splice_reconstructed) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3190
- F1 Score: 0.8937
- Accuracy: 0.8932
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.8107 | 0.7 | 200 | 0.5162 | 0.7800 | 0.7797 |
| 0.4648 | 1.4 | 400 | 0.4492 | 0.8111 | 0.8102 |
| 0.4151 | 2.1 | 600 | 0.4233 | 0.8234 | 0.8224 |
| 0.3773 | 2.8 | 800 | 0.4041 | 0.8388 | 0.8380 |
| 0.3609 | 3.5 | 1000 | 0.3544 | 0.8657 | 0.8652 |
| 0.3422 | 4.2 | 1200 | 0.3838 | 0.8472 | 0.8461 |
| 0.3319 | 4.9 | 1400 | 0.3474 | 0.8673 | 0.8667 |
| 0.3188 | 5.59 | 1600 | 0.3620 | 0.8613 | 0.8608 |
| 0.3061 | 6.29 | 1800 | 0.3711 | 0.8579 | 0.8571 |
| 0.2981 | 6.99 | 2000 | 0.3474 | 0.8684 | 0.8676 |
| 0.288 | 7.69 | 2200 | 0.3728 | 0.8566 | 0.8558 |
| 0.2801 | 8.39 | 2400 | 0.3401 | 0.8776 | 0.8770 |
| 0.2747 | 9.09 | 2600 | 0.3608 | 0.8657 | 0.8645 |
| 0.2667 | 9.79 | 2800 | 0.3399 | 0.8738 | 0.8731 |
| 0.2639 | 10.49 | 3000 | 0.3421 | 0.8739 | 0.8733 |
| 0.2536 | 11.19 | 3200 | 0.3573 | 0.8700 | 0.8694 |
| 0.2512 | 11.89 | 3400 | 0.3525 | 0.8709 | 0.8702 |
| 0.2441 | 12.59 | 3600 | 0.3636 | 0.8666 | 0.8658 |
| 0.2424 | 13.29 | 3800 | 0.3331 | 0.8800 | 0.8794 |
| 0.2365 | 13.99 | 4000 | 0.3711 | 0.8632 | 0.8623 |
| 0.2232 | 14.69 | 4200 | 0.3295 | 0.8784 | 0.8777 |
| 0.2306 | 15.38 | 4400 | 0.3298 | 0.8854 | 0.8849 |
| 0.2251 | 16.08 | 4600 | 0.3545 | 0.8805 | 0.8799 |
| 0.2195 | 16.78 | 4800 | 0.3541 | 0.8745 | 0.8737 |
| 0.2093 | 17.48 | 5000 | 0.3747 | 0.8663 | 0.8654 |
| 0.2134 | 18.18 | 5200 | 0.3485 | 0.8809 | 0.8803 |
| 0.2057 | 18.88 | 5400 | 0.3436 | 0.8848 | 0.8843 |
| 0.1967 | 19.58 | 5600 | 0.3440 | 0.8834 | 0.8829 |
| 0.2058 | 20.28 | 5800 | 0.3337 | 0.8865 | 0.8860 |
| 0.2019 | 20.98 | 6000 | 0.3200 | 0.8899 | 0.8895 |
| 0.1929 | 21.68 | 6200 | 0.3476 | 0.8790 | 0.8783 |
| 0.1886 | 22.38 | 6400 | 0.3352 | 0.8857 | 0.8851 |
| 0.1844 | 23.08 | 6600 | 0.3475 | 0.8827 | 0.8821 |
| 0.1873 | 23.78 | 6800 | 0.3293 | 0.8904 | 0.8900 |
| 0.183 | 24.48 | 7000 | 0.3461 | 0.8822 | 0.8816 |
| 0.1786 | 25.17 | 7200 | 0.3546 | 0.8816 | 0.8810 |
| 0.1799 | 25.87 | 7400 | 0.3289 | 0.8916 | 0.8913 |
| 0.1772 | 26.57 | 7600 | 0.3455 | 0.8864 | 0.8858 |
| 0.1733 | 27.27 | 7800 | 0.3422 | 0.8871 | 0.8867 |
| 0.172 | 27.97 | 8000 | 0.3441 | 0.8902 | 0.8897 |
| 0.1735 | 28.67 | 8200 | 0.3576 | 0.8827 | 0.8821 |
| 0.1665 | 29.37 | 8400 | 0.3407 | 0.8898 | 0.8893 |
| 0.1634 | 30.07 | 8600 | 0.3403 | 0.8932 | 0.8928 |
| 0.1669 | 30.77 | 8800 | 0.3538 | 0.8864 | 0.8858 |
| 0.1638 | 31.47 | 9000 | 0.3552 | 0.8861 | 0.8856 |
| 0.1626 | 32.17 | 9200 | 0.3491 | 0.8898 | 0.8893 |
| 0.1596 | 32.87 | 9400 | 0.3473 | 0.8922 | 0.8917 |
| 0.1587 | 33.57 | 9600 | 0.3634 | 0.8842 | 0.8836 |
| 0.1607 | 34.27 | 9800 | 0.3523 | 0.8904 | 0.8900 |
| 0.1548 | 34.97 | 10000 | 0.3538 | 0.8894 | 0.8889 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_splice_reconstructed-seqsight_16384_512_56M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_splice_reconstructed-seqsight_16384_512_56M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:19:52+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_0-seqsight_16384_512_56M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_0](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_0) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4064
- F1 Score: 0.8127
- Accuracy: 0.814
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5279 | 0.79 | 200 | 0.5073 | 0.7569 | 0.757 |
| 0.489 | 1.58 | 400 | 0.4822 | 0.7810 | 0.781 |
| 0.4803 | 2.37 | 600 | 0.4796 | 0.7750 | 0.775 |
| 0.4718 | 3.16 | 800 | 0.4822 | 0.7689 | 0.769 |
| 0.4662 | 3.95 | 1000 | 0.4704 | 0.7801 | 0.78 |
| 0.4631 | 4.74 | 1200 | 0.4784 | 0.7750 | 0.775 |
| 0.4624 | 5.53 | 1400 | 0.4808 | 0.7752 | 0.776 |
| 0.4561 | 6.32 | 1600 | 0.4741 | 0.7778 | 0.778 |
| 0.4548 | 7.11 | 1800 | 0.4797 | 0.7694 | 0.77 |
| 0.4569 | 7.91 | 2000 | 0.4793 | 0.7694 | 0.77 |
| 0.4524 | 8.7 | 2200 | 0.4693 | 0.7689 | 0.769 |
| 0.451 | 9.49 | 2400 | 0.4635 | 0.7701 | 0.77 |
| 0.4505 | 10.28 | 2600 | 0.4777 | 0.7601 | 0.761 |
| 0.4466 | 11.07 | 2800 | 0.4735 | 0.7713 | 0.772 |
| 0.4467 | 11.86 | 3000 | 0.4717 | 0.7615 | 0.762 |
| 0.4453 | 12.65 | 3200 | 0.4667 | 0.7709 | 0.771 |
| 0.442 | 13.44 | 3400 | 0.4628 | 0.7671 | 0.767 |
| 0.441 | 14.23 | 3600 | 0.4622 | 0.7710 | 0.771 |
| 0.4435 | 15.02 | 3800 | 0.4725 | 0.7584 | 0.759 |
| 0.4419 | 15.81 | 4000 | 0.4650 | 0.7677 | 0.768 |
| 0.4377 | 16.6 | 4200 | 0.4657 | 0.7650 | 0.765 |
| 0.4387 | 17.39 | 4400 | 0.4801 | 0.7598 | 0.761 |
| 0.4395 | 18.18 | 4600 | 0.4644 | 0.7609 | 0.761 |
| 0.4367 | 18.97 | 4800 | 0.4738 | 0.7650 | 0.766 |
| 0.439 | 19.76 | 5000 | 0.4657 | 0.7658 | 0.766 |
| 0.4355 | 20.55 | 5200 | 0.4583 | 0.7681 | 0.768 |
| 0.4328 | 21.34 | 5400 | 0.4705 | 0.7625 | 0.763 |
| 0.4285 | 22.13 | 5600 | 0.4681 | 0.7670 | 0.767 |
| 0.4361 | 22.92 | 5800 | 0.4673 | 0.7679 | 0.768 |
| 0.427 | 23.72 | 6000 | 0.4720 | 0.7645 | 0.765 |
| 0.4315 | 24.51 | 6200 | 0.4585 | 0.7711 | 0.771 |
| 0.4353 | 25.3 | 6400 | 0.4647 | 0.7660 | 0.766 |
| 0.4273 | 26.09 | 6600 | 0.4611 | 0.7681 | 0.768 |
| 0.4285 | 26.88 | 6800 | 0.4691 | 0.7698 | 0.77 |
| 0.4284 | 27.67 | 7000 | 0.4606 | 0.7701 | 0.77 |
| 0.4303 | 28.46 | 7200 | 0.4568 | 0.7691 | 0.769 |
| 0.4264 | 29.25 | 7400 | 0.4654 | 0.7750 | 0.775 |
| 0.4299 | 30.04 | 7600 | 0.4647 | 0.7680 | 0.768 |
| 0.4272 | 30.83 | 7800 | 0.4631 | 0.7700 | 0.77 |
| 0.4281 | 31.62 | 8000 | 0.4580 | 0.7701 | 0.77 |
| 0.4273 | 32.41 | 8200 | 0.4676 | 0.7689 | 0.769 |
| 0.423 | 33.2 | 8400 | 0.4669 | 0.77 | 0.77 |
| 0.4275 | 33.99 | 8600 | 0.4635 | 0.7720 | 0.772 |
| 0.4269 | 34.78 | 8800 | 0.4598 | 0.7721 | 0.772 |
| 0.4217 | 35.57 | 9000 | 0.4624 | 0.7700 | 0.77 |
| 0.4234 | 36.36 | 9200 | 0.4682 | 0.7679 | 0.768 |
| 0.4268 | 37.15 | 9400 | 0.4682 | 0.7679 | 0.768 |
| 0.423 | 37.94 | 9600 | 0.4634 | 0.7700 | 0.77 |
| 0.4262 | 38.74 | 9800 | 0.4653 | 0.77 | 0.77 |
| 0.4253 | 39.53 | 10000 | 0.4645 | 0.7720 | 0.772 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_0-seqsight_16384_512_56M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_0-seqsight_16384_512_56M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:20:30+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_0-seqsight_16384_512_56M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_0](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_0) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3572
- F1 Score: 0.8419
- Accuracy: 0.842
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5131 | 0.79 | 200 | 0.4878 | 0.7721 | 0.772 |
| 0.476 | 1.58 | 400 | 0.4730 | 0.7811 | 0.781 |
| 0.4664 | 2.37 | 600 | 0.4650 | 0.7749 | 0.775 |
| 0.4582 | 3.16 | 800 | 0.4656 | 0.7730 | 0.773 |
| 0.4531 | 3.95 | 1000 | 0.4592 | 0.7731 | 0.773 |
| 0.4503 | 4.74 | 1200 | 0.4637 | 0.7751 | 0.775 |
| 0.4464 | 5.53 | 1400 | 0.4710 | 0.7663 | 0.767 |
| 0.4406 | 6.32 | 1600 | 0.4647 | 0.7739 | 0.774 |
| 0.4399 | 7.11 | 1800 | 0.4696 | 0.7768 | 0.777 |
| 0.4398 | 7.91 | 2000 | 0.4667 | 0.7730 | 0.773 |
| 0.4334 | 8.7 | 2200 | 0.4690 | 0.7669 | 0.767 |
| 0.4311 | 9.49 | 2400 | 0.4585 | 0.7790 | 0.779 |
| 0.4331 | 10.28 | 2600 | 0.4673 | 0.7709 | 0.771 |
| 0.4267 | 11.07 | 2800 | 0.4655 | 0.7779 | 0.778 |
| 0.4263 | 11.86 | 3000 | 0.4720 | 0.7750 | 0.775 |
| 0.4228 | 12.65 | 3200 | 0.4595 | 0.7761 | 0.776 |
| 0.4199 | 13.44 | 3400 | 0.4621 | 0.7691 | 0.769 |
| 0.4174 | 14.23 | 3600 | 0.4661 | 0.7751 | 0.775 |
| 0.4202 | 15.02 | 3800 | 0.4801 | 0.7668 | 0.767 |
| 0.4168 | 15.81 | 4000 | 0.4601 | 0.7671 | 0.767 |
| 0.4133 | 16.6 | 4200 | 0.4763 | 0.7720 | 0.772 |
| 0.4158 | 17.39 | 4400 | 0.4739 | 0.7729 | 0.773 |
| 0.4121 | 18.18 | 4600 | 0.4648 | 0.7739 | 0.774 |
| 0.4107 | 18.97 | 4800 | 0.4741 | 0.7690 | 0.769 |
| 0.4117 | 19.76 | 5000 | 0.4676 | 0.7670 | 0.767 |
| 0.4075 | 20.55 | 5200 | 0.4679 | 0.7710 | 0.771 |
| 0.4064 | 21.34 | 5400 | 0.4830 | 0.7796 | 0.78 |
| 0.4025 | 22.13 | 5600 | 0.4792 | 0.7761 | 0.776 |
| 0.4084 | 22.92 | 5800 | 0.4737 | 0.7781 | 0.778 |
| 0.3996 | 23.72 | 6000 | 0.4787 | 0.7789 | 0.779 |
| 0.4024 | 24.51 | 6200 | 0.4653 | 0.7790 | 0.779 |
| 0.4052 | 25.3 | 6400 | 0.4668 | 0.7771 | 0.777 |
| 0.3962 | 26.09 | 6600 | 0.4679 | 0.7788 | 0.779 |
| 0.4 | 26.88 | 6800 | 0.4761 | 0.7781 | 0.778 |
| 0.3991 | 27.67 | 7000 | 0.4695 | 0.7779 | 0.778 |
| 0.3986 | 28.46 | 7200 | 0.4698 | 0.7808 | 0.781 |
| 0.3974 | 29.25 | 7400 | 0.4747 | 0.7771 | 0.777 |
| 0.3999 | 30.04 | 7600 | 0.4768 | 0.7810 | 0.781 |
| 0.3937 | 30.83 | 7800 | 0.4715 | 0.7800 | 0.78 |
| 0.3957 | 31.62 | 8000 | 0.4713 | 0.7799 | 0.78 |
| 0.398 | 32.41 | 8200 | 0.4757 | 0.7791 | 0.779 |
| 0.3889 | 33.2 | 8400 | 0.4788 | 0.7811 | 0.781 |
| 0.3946 | 33.99 | 8600 | 0.4763 | 0.7801 | 0.78 |
| 0.3939 | 34.78 | 8800 | 0.4716 | 0.7789 | 0.779 |
| 0.3882 | 35.57 | 9000 | 0.4757 | 0.7770 | 0.777 |
| 0.3912 | 36.36 | 9200 | 0.4801 | 0.7841 | 0.784 |
| 0.3912 | 37.15 | 9400 | 0.4793 | 0.7830 | 0.783 |
| 0.3917 | 37.94 | 9600 | 0.4744 | 0.7801 | 0.78 |
| 0.3929 | 38.74 | 9800 | 0.4754 | 0.7811 | 0.781 |
| 0.3907 | 39.53 | 10000 | 0.4752 | 0.7801 | 0.78 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_0-seqsight_16384_512_56M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_0-seqsight_16384_512_56M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:21:00+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_0-seqsight_16384_512_56M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_0](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_0) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3661
- F1 Score: 0.8357
- Accuracy: 0.836
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5029 | 0.79 | 200 | 0.4775 | 0.7770 | 0.777 |
| 0.469 | 1.58 | 400 | 0.4746 | 0.7719 | 0.772 |
| 0.4599 | 2.37 | 600 | 0.4616 | 0.7796 | 0.781 |
| 0.4515 | 3.16 | 800 | 0.4692 | 0.7691 | 0.769 |
| 0.4449 | 3.95 | 1000 | 0.4522 | 0.7798 | 0.78 |
| 0.4415 | 4.74 | 1200 | 0.4581 | 0.7831 | 0.783 |
| 0.4331 | 5.53 | 1400 | 0.4776 | 0.7612 | 0.762 |
| 0.4282 | 6.32 | 1600 | 0.4613 | 0.7821 | 0.782 |
| 0.4264 | 7.11 | 1800 | 0.4759 | 0.7719 | 0.772 |
| 0.4251 | 7.91 | 2000 | 0.4792 | 0.7748 | 0.775 |
| 0.4151 | 8.7 | 2200 | 0.4855 | 0.7671 | 0.767 |
| 0.4126 | 9.49 | 2400 | 0.4732 | 0.7730 | 0.773 |
| 0.4133 | 10.28 | 2600 | 0.4940 | 0.7675 | 0.768 |
| 0.4054 | 11.07 | 2800 | 0.4728 | 0.7651 | 0.765 |
| 0.4034 | 11.86 | 3000 | 0.4815 | 0.7771 | 0.777 |
| 0.3966 | 12.65 | 3200 | 0.4756 | 0.7811 | 0.781 |
| 0.3922 | 13.44 | 3400 | 0.4889 | 0.7760 | 0.776 |
| 0.3908 | 14.23 | 3600 | 0.4940 | 0.7681 | 0.768 |
| 0.3915 | 15.02 | 3800 | 0.4998 | 0.7769 | 0.777 |
| 0.3847 | 15.81 | 4000 | 0.4762 | 0.7711 | 0.771 |
| 0.381 | 16.6 | 4200 | 0.5055 | 0.7690 | 0.769 |
| 0.3819 | 17.39 | 4400 | 0.4960 | 0.7748 | 0.775 |
| 0.3754 | 18.18 | 4600 | 0.4943 | 0.7750 | 0.775 |
| 0.3719 | 18.97 | 4800 | 0.5116 | 0.7590 | 0.759 |
| 0.3708 | 19.76 | 5000 | 0.4967 | 0.7640 | 0.764 |
| 0.3669 | 20.55 | 5200 | 0.5051 | 0.7711 | 0.771 |
| 0.3638 | 21.34 | 5400 | 0.5155 | 0.7607 | 0.761 |
| 0.3586 | 22.13 | 5600 | 0.5215 | 0.7619 | 0.762 |
| 0.3618 | 22.92 | 5800 | 0.5110 | 0.7638 | 0.764 |
| 0.3558 | 23.72 | 6000 | 0.5217 | 0.7627 | 0.763 |
| 0.3545 | 24.51 | 6200 | 0.5119 | 0.7600 | 0.76 |
| 0.3558 | 25.3 | 6400 | 0.5195 | 0.7649 | 0.765 |
| 0.3461 | 26.09 | 6600 | 0.5263 | 0.7731 | 0.773 |
| 0.3482 | 26.88 | 6800 | 0.5298 | 0.7601 | 0.76 |
| 0.3463 | 27.67 | 7000 | 0.5299 | 0.7651 | 0.765 |
| 0.3429 | 28.46 | 7200 | 0.5497 | 0.7571 | 0.757 |
| 0.342 | 29.25 | 7400 | 0.5309 | 0.7661 | 0.766 |
| 0.3407 | 30.04 | 7600 | 0.5545 | 0.7625 | 0.763 |
| 0.3352 | 30.83 | 7800 | 0.5369 | 0.7671 | 0.767 |
| 0.3357 | 31.62 | 8000 | 0.5440 | 0.7591 | 0.759 |
| 0.3344 | 32.41 | 8200 | 0.5570 | 0.7569 | 0.757 |
| 0.3274 | 33.2 | 8400 | 0.5703 | 0.7629 | 0.763 |
| 0.3339 | 33.99 | 8600 | 0.5548 | 0.7670 | 0.767 |
| 0.3317 | 34.78 | 8800 | 0.5468 | 0.7721 | 0.772 |
| 0.3244 | 35.57 | 9000 | 0.5577 | 0.7671 | 0.767 |
| 0.3266 | 36.36 | 9200 | 0.5741 | 0.7688 | 0.769 |
| 0.3245 | 37.15 | 9400 | 0.5671 | 0.7668 | 0.767 |
| 0.3274 | 37.94 | 9600 | 0.5610 | 0.7639 | 0.764 |
| 0.3255 | 38.74 | 9800 | 0.5624 | 0.7679 | 0.768 |
| 0.3241 | 39.53 | 10000 | 0.5632 | 0.7670 | 0.767 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_0-seqsight_16384_512_56M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_0-seqsight_16384_512_56M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:21:28+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_1-seqsight_16384_512_56M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_1](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_1) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3366
- F1 Score: 0.8579
- Accuracy: 0.858
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.551 | 0.83 | 200 | 0.5506 | 0.7238 | 0.724 |
| 0.5064 | 1.67 | 400 | 0.5437 | 0.7329 | 0.733 |
| 0.4977 | 2.5 | 600 | 0.5557 | 0.7164 | 0.719 |
| 0.4899 | 3.33 | 800 | 0.5347 | 0.7231 | 0.724 |
| 0.4908 | 4.17 | 1000 | 0.5390 | 0.7245 | 0.726 |
| 0.4844 | 5.0 | 1200 | 0.5362 | 0.7322 | 0.733 |
| 0.4799 | 5.83 | 1400 | 0.5291 | 0.7330 | 0.734 |
| 0.4822 | 6.67 | 1600 | 0.5358 | 0.7273 | 0.729 |
| 0.4762 | 7.5 | 1800 | 0.5175 | 0.7479 | 0.748 |
| 0.4722 | 8.33 | 2000 | 0.5286 | 0.7323 | 0.734 |
| 0.4729 | 9.17 | 2200 | 0.5253 | 0.7385 | 0.74 |
| 0.4711 | 10.0 | 2400 | 0.5386 | 0.7332 | 0.737 |
| 0.4699 | 10.83 | 2600 | 0.5331 | 0.7392 | 0.743 |
| 0.468 | 11.67 | 2800 | 0.5377 | 0.7285 | 0.733 |
| 0.461 | 12.5 | 3000 | 0.5235 | 0.7438 | 0.746 |
| 0.4719 | 13.33 | 3200 | 0.5198 | 0.7397 | 0.743 |
| 0.4633 | 14.17 | 3400 | 0.5141 | 0.7414 | 0.742 |
| 0.4617 | 15.0 | 3600 | 0.5352 | 0.7354 | 0.74 |
| 0.4591 | 15.83 | 3800 | 0.5175 | 0.7398 | 0.742 |
| 0.4637 | 16.67 | 4000 | 0.5278 | 0.7374 | 0.74 |
| 0.4575 | 17.5 | 4200 | 0.5248 | 0.7407 | 0.744 |
| 0.4553 | 18.33 | 4400 | 0.5220 | 0.7466 | 0.748 |
| 0.459 | 19.17 | 4600 | 0.5385 | 0.7329 | 0.738 |
| 0.4594 | 20.0 | 4800 | 0.5129 | 0.7516 | 0.753 |
| 0.4578 | 20.83 | 5000 | 0.5249 | 0.7446 | 0.747 |
| 0.4526 | 21.67 | 5200 | 0.5175 | 0.7502 | 0.752 |
| 0.4504 | 22.5 | 5400 | 0.5147 | 0.7510 | 0.752 |
| 0.4526 | 23.33 | 5600 | 0.5190 | 0.7457 | 0.748 |
| 0.4516 | 24.17 | 5800 | 0.5322 | 0.7434 | 0.747 |
| 0.4515 | 25.0 | 6000 | 0.5098 | 0.7523 | 0.753 |
| 0.4489 | 25.83 | 6200 | 0.5128 | 0.7412 | 0.743 |
| 0.4477 | 26.67 | 6400 | 0.5093 | 0.7544 | 0.755 |
| 0.4512 | 27.5 | 6600 | 0.5186 | 0.7474 | 0.749 |
| 0.4477 | 28.33 | 6800 | 0.5134 | 0.7464 | 0.748 |
| 0.4467 | 29.17 | 7000 | 0.5177 | 0.7415 | 0.743 |
| 0.448 | 30.0 | 7200 | 0.5153 | 0.7420 | 0.744 |
| 0.4472 | 30.83 | 7400 | 0.5255 | 0.7382 | 0.742 |
| 0.4451 | 31.67 | 7600 | 0.5124 | 0.7496 | 0.751 |
| 0.445 | 32.5 | 7800 | 0.5146 | 0.7526 | 0.754 |
| 0.4428 | 33.33 | 8000 | 0.5151 | 0.7425 | 0.744 |
| 0.4477 | 34.17 | 8200 | 0.5125 | 0.7556 | 0.757 |
| 0.4437 | 35.0 | 8400 | 0.5093 | 0.7520 | 0.753 |
| 0.4459 | 35.83 | 8600 | 0.5098 | 0.7536 | 0.755 |
| 0.446 | 36.67 | 8800 | 0.5121 | 0.7512 | 0.753 |
| 0.4422 | 37.5 | 9000 | 0.5227 | 0.7405 | 0.744 |
| 0.4419 | 38.33 | 9200 | 0.5149 | 0.7481 | 0.75 |
| 0.4424 | 39.17 | 9400 | 0.5133 | 0.7503 | 0.752 |
| 0.4425 | 40.0 | 9600 | 0.5128 | 0.7513 | 0.753 |
| 0.4418 | 40.83 | 9800 | 0.5110 | 0.7550 | 0.756 |
| 0.4447 | 41.67 | 10000 | 0.5125 | 0.7514 | 0.753 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_1-seqsight_16384_512_56M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_1-seqsight_16384_512_56M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:21:40+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_1-seqsight_16384_512_56M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_1](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_1) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3495
- F1 Score: 0.8466
- Accuracy: 0.847
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5368 | 0.83 | 200 | 0.5383 | 0.7306 | 0.731 |
| 0.4923 | 1.67 | 400 | 0.5293 | 0.7299 | 0.73 |
| 0.483 | 2.5 | 600 | 0.5362 | 0.7371 | 0.739 |
| 0.4758 | 3.33 | 800 | 0.5158 | 0.7378 | 0.738 |
| 0.4746 | 4.17 | 1000 | 0.5274 | 0.7311 | 0.733 |
| 0.4688 | 5.0 | 1200 | 0.5218 | 0.7352 | 0.737 |
| 0.4637 | 5.83 | 1400 | 0.5080 | 0.7336 | 0.734 |
| 0.4634 | 6.67 | 1600 | 0.5112 | 0.7426 | 0.743 |
| 0.4601 | 7.5 | 1800 | 0.5006 | 0.752 | 0.752 |
| 0.4516 | 8.33 | 2000 | 0.5084 | 0.7438 | 0.744 |
| 0.4533 | 9.17 | 2200 | 0.5005 | 0.7468 | 0.747 |
| 0.4492 | 10.0 | 2400 | 0.5322 | 0.7374 | 0.74 |
| 0.4471 | 10.83 | 2600 | 0.5154 | 0.7357 | 0.738 |
| 0.4448 | 11.67 | 2800 | 0.5192 | 0.7329 | 0.735 |
| 0.4376 | 12.5 | 3000 | 0.5148 | 0.7440 | 0.745 |
| 0.4477 | 13.33 | 3200 | 0.5106 | 0.7415 | 0.743 |
| 0.4389 | 14.17 | 3400 | 0.5060 | 0.7530 | 0.753 |
| 0.4372 | 15.0 | 3600 | 0.5154 | 0.7426 | 0.744 |
| 0.4334 | 15.83 | 3800 | 0.5093 | 0.7404 | 0.742 |
| 0.4374 | 16.67 | 4000 | 0.5132 | 0.7396 | 0.74 |
| 0.4322 | 17.5 | 4200 | 0.5190 | 0.7325 | 0.735 |
| 0.4262 | 18.33 | 4400 | 0.5159 | 0.7506 | 0.751 |
| 0.432 | 19.17 | 4600 | 0.5306 | 0.7279 | 0.73 |
| 0.4319 | 20.0 | 4800 | 0.5069 | 0.7528 | 0.753 |
| 0.4289 | 20.83 | 5000 | 0.5240 | 0.7434 | 0.744 |
| 0.4237 | 21.67 | 5200 | 0.5160 | 0.7411 | 0.742 |
| 0.422 | 22.5 | 5400 | 0.5161 | 0.7447 | 0.745 |
| 0.4229 | 23.33 | 5600 | 0.5230 | 0.7362 | 0.738 |
| 0.4227 | 24.17 | 5800 | 0.5297 | 0.7292 | 0.731 |
| 0.4221 | 25.0 | 6000 | 0.5119 | 0.7550 | 0.755 |
| 0.4193 | 25.83 | 6200 | 0.5071 | 0.7496 | 0.75 |
| 0.4172 | 26.67 | 6400 | 0.5120 | 0.7540 | 0.754 |
| 0.4204 | 27.5 | 6600 | 0.5207 | 0.7488 | 0.749 |
| 0.4151 | 28.33 | 6800 | 0.5089 | 0.7490 | 0.749 |
| 0.4157 | 29.17 | 7000 | 0.5147 | 0.7527 | 0.753 |
| 0.4146 | 30.0 | 7200 | 0.5101 | 0.7468 | 0.747 |
| 0.415 | 30.83 | 7400 | 0.5164 | 0.7362 | 0.737 |
| 0.4117 | 31.67 | 7600 | 0.5122 | 0.7468 | 0.747 |
| 0.4127 | 32.5 | 7800 | 0.5193 | 0.7518 | 0.752 |
| 0.4117 | 33.33 | 8000 | 0.5125 | 0.7477 | 0.748 |
| 0.4127 | 34.17 | 8200 | 0.5173 | 0.7446 | 0.745 |
| 0.4095 | 35.0 | 8400 | 0.5108 | 0.7519 | 0.752 |
| 0.4111 | 35.83 | 8600 | 0.5113 | 0.7468 | 0.747 |
| 0.4122 | 36.67 | 8800 | 0.5123 | 0.7457 | 0.746 |
| 0.4073 | 37.5 | 9000 | 0.5191 | 0.7409 | 0.742 |
| 0.4063 | 38.33 | 9200 | 0.5144 | 0.7448 | 0.745 |
| 0.4076 | 39.17 | 9400 | 0.5159 | 0.7468 | 0.747 |
| 0.4071 | 40.0 | 9600 | 0.5162 | 0.7468 | 0.747 |
| 0.4073 | 40.83 | 9800 | 0.5144 | 0.7469 | 0.747 |
| 0.4078 | 41.67 | 10000 | 0.5144 | 0.7488 | 0.749 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_1-seqsight_16384_512_56M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_1-seqsight_16384_512_56M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:21:51+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_1-seqsight_16384_512_56M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_1](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_1) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3980
- F1 Score: 0.8258
- Accuracy: 0.826
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5287 | 0.83 | 200 | 0.5321 | 0.7324 | 0.733 |
| 0.487 | 1.67 | 400 | 0.5212 | 0.7370 | 0.737 |
| 0.4764 | 2.5 | 600 | 0.5245 | 0.7351 | 0.737 |
| 0.4678 | 3.33 | 800 | 0.5001 | 0.7326 | 0.733 |
| 0.464 | 4.17 | 1000 | 0.5146 | 0.7270 | 0.729 |
| 0.4565 | 5.0 | 1200 | 0.5174 | 0.7307 | 0.733 |
| 0.4507 | 5.83 | 1400 | 0.5056 | 0.7335 | 0.734 |
| 0.4486 | 6.67 | 1600 | 0.5054 | 0.7427 | 0.743 |
| 0.445 | 7.5 | 1800 | 0.5034 | 0.7478 | 0.748 |
| 0.4353 | 8.33 | 2000 | 0.5063 | 0.7437 | 0.744 |
| 0.4378 | 9.17 | 2200 | 0.5070 | 0.7477 | 0.748 |
| 0.432 | 10.0 | 2400 | 0.5209 | 0.7395 | 0.741 |
| 0.4291 | 10.83 | 2600 | 0.5171 | 0.7413 | 0.743 |
| 0.4236 | 11.67 | 2800 | 0.5239 | 0.7356 | 0.737 |
| 0.4163 | 12.5 | 3000 | 0.5192 | 0.7430 | 0.744 |
| 0.4232 | 13.33 | 3200 | 0.5140 | 0.7420 | 0.743 |
| 0.414 | 14.17 | 3400 | 0.5150 | 0.7460 | 0.746 |
| 0.4118 | 15.0 | 3600 | 0.5221 | 0.7284 | 0.73 |
| 0.4061 | 15.83 | 3800 | 0.5109 | 0.7376 | 0.739 |
| 0.4062 | 16.67 | 4000 | 0.5263 | 0.7392 | 0.74 |
| 0.4005 | 17.5 | 4200 | 0.5359 | 0.7326 | 0.735 |
| 0.393 | 18.33 | 4400 | 0.5274 | 0.7392 | 0.74 |
| 0.3957 | 19.17 | 4600 | 0.5527 | 0.7341 | 0.736 |
| 0.3951 | 20.0 | 4800 | 0.5196 | 0.7430 | 0.743 |
| 0.3885 | 20.83 | 5000 | 0.5419 | 0.7287 | 0.73 |
| 0.3836 | 21.67 | 5200 | 0.5337 | 0.7449 | 0.746 |
| 0.3793 | 22.5 | 5400 | 0.5479 | 0.7318 | 0.733 |
| 0.3784 | 23.33 | 5600 | 0.5471 | 0.7304 | 0.733 |
| 0.378 | 24.17 | 5800 | 0.5490 | 0.7218 | 0.724 |
| 0.3738 | 25.0 | 6000 | 0.5380 | 0.7468 | 0.747 |
| 0.3718 | 25.83 | 6200 | 0.5386 | 0.7366 | 0.737 |
| 0.3666 | 26.67 | 6400 | 0.5330 | 0.7559 | 0.756 |
| 0.3681 | 27.5 | 6600 | 0.5577 | 0.7447 | 0.745 |
| 0.3655 | 28.33 | 6800 | 0.5525 | 0.7510 | 0.751 |
| 0.3623 | 29.17 | 7000 | 0.5424 | 0.7487 | 0.749 |
| 0.3603 | 30.0 | 7200 | 0.5409 | 0.7358 | 0.736 |
| 0.3593 | 30.83 | 7400 | 0.5560 | 0.7321 | 0.733 |
| 0.3543 | 31.67 | 7600 | 0.5555 | 0.7358 | 0.736 |
| 0.3546 | 32.5 | 7800 | 0.5665 | 0.7409 | 0.741 |
| 0.3521 | 33.33 | 8000 | 0.5579 | 0.7326 | 0.733 |
| 0.3502 | 34.17 | 8200 | 0.5709 | 0.7426 | 0.743 |
| 0.3483 | 35.0 | 8400 | 0.5577 | 0.7447 | 0.745 |
| 0.3487 | 35.83 | 8600 | 0.5595 | 0.7426 | 0.743 |
| 0.3474 | 36.67 | 8800 | 0.5614 | 0.7405 | 0.741 |
| 0.3431 | 37.5 | 9000 | 0.5769 | 0.7352 | 0.736 |
| 0.3433 | 38.33 | 9200 | 0.5750 | 0.7365 | 0.737 |
| 0.3426 | 39.17 | 9400 | 0.5791 | 0.7346 | 0.735 |
| 0.3392 | 40.0 | 9600 | 0.5810 | 0.7345 | 0.735 |
| 0.3392 | 40.83 | 9800 | 0.5783 | 0.7378 | 0.738 |
| 0.3414 | 41.67 | 10000 | 0.5779 | 0.7316 | 0.732 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_1-seqsight_16384_512_56M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_1-seqsight_16384_512_56M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:22:11+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_4-seqsight_16384_512_56M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_4](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_4) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3561
- F1 Score: 0.8409
- Accuracy: 0.841
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5395 | 1.34 | 200 | 0.5284 | 0.7419 | 0.745 |
| 0.486 | 2.68 | 400 | 0.4936 | 0.7544 | 0.755 |
| 0.4753 | 4.03 | 600 | 0.4808 | 0.7710 | 0.771 |
| 0.4613 | 5.37 | 800 | 0.4943 | 0.7670 | 0.768 |
| 0.4526 | 6.71 | 1000 | 0.4694 | 0.7810 | 0.781 |
| 0.4474 | 8.05 | 1200 | 0.4778 | 0.7761 | 0.777 |
| 0.4436 | 9.4 | 1400 | 0.4758 | 0.7753 | 0.776 |
| 0.4385 | 10.74 | 1600 | 0.4700 | 0.7859 | 0.786 |
| 0.4331 | 12.08 | 1800 | 0.4661 | 0.7810 | 0.781 |
| 0.4312 | 13.42 | 2000 | 0.4612 | 0.7867 | 0.787 |
| 0.4204 | 14.77 | 2200 | 0.4806 | 0.7734 | 0.775 |
| 0.4201 | 16.11 | 2400 | 0.4583 | 0.7909 | 0.791 |
| 0.417 | 17.45 | 2600 | 0.4637 | 0.7791 | 0.78 |
| 0.4147 | 18.79 | 2800 | 0.4676 | 0.7771 | 0.778 |
| 0.4117 | 20.13 | 3000 | 0.4746 | 0.7748 | 0.777 |
| 0.4091 | 21.48 | 3200 | 0.4557 | 0.7887 | 0.789 |
| 0.4041 | 22.82 | 3400 | 0.4532 | 0.7870 | 0.787 |
| 0.4028 | 24.16 | 3600 | 0.4506 | 0.7929 | 0.793 |
| 0.4015 | 25.5 | 3800 | 0.4427 | 0.7900 | 0.79 |
| 0.3975 | 26.85 | 4000 | 0.4468 | 0.7958 | 0.796 |
| 0.3963 | 28.19 | 4200 | 0.4466 | 0.7939 | 0.794 |
| 0.3913 | 29.53 | 4400 | 0.4700 | 0.7737 | 0.775 |
| 0.3942 | 30.87 | 4600 | 0.4518 | 0.7946 | 0.795 |
| 0.389 | 32.21 | 4800 | 0.4467 | 0.7967 | 0.797 |
| 0.3857 | 33.56 | 5000 | 0.4608 | 0.7885 | 0.789 |
| 0.3896 | 34.9 | 5200 | 0.4592 | 0.7864 | 0.787 |
| 0.3893 | 36.24 | 5400 | 0.4456 | 0.7978 | 0.798 |
| 0.3816 | 37.58 | 5600 | 0.4475 | 0.7989 | 0.799 |
| 0.3821 | 38.93 | 5800 | 0.4647 | 0.7818 | 0.783 |
| 0.3802 | 40.27 | 6000 | 0.4459 | 0.8028 | 0.803 |
| 0.3796 | 41.61 | 6200 | 0.4573 | 0.7872 | 0.788 |
| 0.3807 | 42.95 | 6400 | 0.4567 | 0.7903 | 0.791 |
| 0.3804 | 44.3 | 6600 | 0.4409 | 0.8018 | 0.802 |
| 0.3721 | 45.64 | 6800 | 0.4573 | 0.7901 | 0.791 |
| 0.3791 | 46.98 | 7000 | 0.4564 | 0.7871 | 0.788 |
| 0.376 | 48.32 | 7200 | 0.4486 | 0.7986 | 0.799 |
| 0.3744 | 49.66 | 7400 | 0.4501 | 0.8005 | 0.801 |
| 0.3718 | 51.01 | 7600 | 0.4491 | 0.7934 | 0.794 |
| 0.3715 | 52.35 | 7800 | 0.4394 | 0.8059 | 0.806 |
| 0.3687 | 53.69 | 8000 | 0.4543 | 0.7943 | 0.795 |
| 0.3709 | 55.03 | 8200 | 0.4594 | 0.7881 | 0.789 |
| 0.371 | 56.38 | 8400 | 0.4488 | 0.8006 | 0.801 |
| 0.3664 | 57.72 | 8600 | 0.4429 | 0.8058 | 0.806 |
| 0.3694 | 59.06 | 8800 | 0.4443 | 0.8037 | 0.804 |
| 0.3679 | 60.4 | 9000 | 0.4477 | 0.8006 | 0.801 |
| 0.3655 | 61.74 | 9200 | 0.4410 | 0.8079 | 0.808 |
| 0.3665 | 63.09 | 9400 | 0.4451 | 0.8037 | 0.804 |
| 0.368 | 64.43 | 9600 | 0.4406 | 0.8068 | 0.807 |
| 0.3686 | 65.77 | 9800 | 0.4448 | 0.8047 | 0.805 |
| 0.3649 | 67.11 | 10000 | 0.4442 | 0.8047 | 0.805 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_4-seqsight_16384_512_56M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_4-seqsight_16384_512_56M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:22:45+00:00 |
text-generation | null |
# kat33/Llama-3-8B-Instruct-Gradient-1048k-Q6_K-GGUF
This model was converted to GGUF format from [`gradientai/Llama-3-8B-Instruct-Gradient-1048k`](https://huggingface.co/gradientai/Llama-3-8B-Instruct-Gradient-1048k) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
Refer to the [original model card](https://huggingface.co/gradientai/Llama-3-8B-Instruct-Gradient-1048k) for more details on the model.
## Use with llama.cpp
Install llama.cpp through brew.
```bash
brew install ggerganov/ggerganov/llama.cpp
```
Invoke the llama.cpp server or the CLI.
CLI:
```bash
llama-cli --hf-repo kat33/Llama-3-8B-Instruct-Gradient-1048k-Q6_K-GGUF --model llama-3-8b-instruct-gradient-1048k.Q6_K.gguf -p "The meaning to life and the universe is"
```
Server:
```bash
llama-server --hf-repo kat33/Llama-3-8B-Instruct-Gradient-1048k-Q6_K-GGUF --model llama-3-8b-instruct-gradient-1048k.Q6_K.gguf -c 2048
```
Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well.
```
git clone https://github.com/ggerganov/llama.cpp && cd llama.cpp && make && ./main -m llama-3-8b-instruct-gradient-1048k.Q6_K.gguf -n 128
```
| {"language": ["en"], "license": "llama3", "tags": ["meta", "llama-3", "llama-cpp", "gguf-my-repo"], "pipeline_tag": "text-generation"} | kat33/Llama-3-8B-Instruct-Gradient-1048k-Q6_K-GGUF | null | [
"gguf",
"meta",
"llama-3",
"llama-cpp",
"gguf-my-repo",
"text-generation",
"en",
"license:llama3",
"region:us"
] | null | 2024-04-30T03:23:45+00:00 |
null | transformers |
<p align="center">
<img src="https://doctr-static.mindee.com/models?id=v0.3.1/Logo_doctr.gif&src=0" width="60%">
</p>
**Optical Character Recognition made seamless & accessible to anyone, powered by TensorFlow 2 & PyTorch**
## Task: detection
https://github.com/mindee/doctr
### Example usage:
```python
>>> from doctr.io import DocumentFile
>>> from doctr.models import ocr_predictor, from_hub
>>> img = DocumentFile.from_images(['<image_path>'])
>>> # Load your model from the hub
>>> model = from_hub('mindee/my-model')
>>> # Pass it to the predictor
>>> # If your model is a recognition model:
>>> predictor = ocr_predictor(det_arch='db_mobilenet_v3_large',
>>> reco_arch=model,
>>> pretrained=True)
>>> # If your model is a detection model:
>>> predictor = ocr_predictor(det_arch=model,
>>> reco_arch='crnn_mobilenet_v3_small',
>>> pretrained=True)
>>> # Get your predictions
>>> res = predictor(img)
```
### Run Configuration
{
"train_path": "/workspace/donut_train/doctr/train/",
"val_path": "/workspace/donut_train/doctr/val/",
"arch": "db_resnet50",
"name": "detection_test",
"epochs": 15,
"batch_size": 2,
"device": 0,
"save_interval_epoch": false,
"input_size": 1024,
"lr": 0.001,
"weight_decay": 0,
"workers": 16,
"resume": null,
"test_only": false,
"freeze_backbone": false,
"show_samples": false,
"wb": true,
"push_to_hub": true,
"pretrained": false,
"rotation": false,
"eval_straight": false,
"sched": "poly",
"amp": false,
"find_lr": false,
"early_stop": false,
"early_stop_epochs": 5,
"early_stop_delta": 0.01
} | {"language": "en"} | Alexleetw/detection_test | null | [
"transformers",
"pytorch",
"en",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:23:59+00:00 |
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | cilantro9246/kd7qkzx | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:24:04+00:00 |
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | shallow6414/yghjnvs | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:24:42+00:00 |
text-generation | transformers | <a href="https://www.gradient.ai" target="_blank"><img src="https://cdn-uploads.huggingface.co/production/uploads/655bb613e8a8971e89944f3e/TSa3V8YpoVagnTYgxiLaO.png" width="200"/></a>
# Llama-3 8B Gradient Instruct 1048k
Gradient incorporates your data to deploy autonomous assistants that power critical operations across your business. If you're looking to build custom AI models or agents, email us a message [email protected].
For more info see our [End-to-end development service for custom LLMs and AI systems](https://gradient.ai/development-lab)
This model extends LLama-3 8B's context length from 8k to > 1040K, developed by Gradient, sponsored by compute from [Crusoe Energy](https://huggingface.co/crusoeai). It demonstrates that SOTA LLMs can learn to operate on long context with minimal training by appropriately adjusting RoPE theta. We trained on 830M tokens for this stage, and 1.4B tokens total for all stages, which is < 0.01% of Llama-3's original pre-training data.
**Approach:**
- [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) as the base
- NTK-aware interpolation [1] to initialize an optimal schedule for RoPE theta, followed by empirical RoPE theta optimization
- Progressive training on increasing context lengths, similar to [Large World Model](https://huggingface.co/LargeWorldModel) [2] (See details below)
**Infra:**
We build on top of the EasyContext Blockwise RingAttention library [3] to scalably and efficiently train on contexts up to 1048k tokens on [Crusoe Energy](https://huggingface.co/crusoeai) high performance L40S cluster.
Notably, we layered parallelism on top of Ring Attention with a custom network topology to better leverage large GPU clusters in the face of network bottlenecks from passing many KV blocks between devices. This gave us a 33x speedup in model training (compare 524k and 1048k to 65k and 262k in the table below).
**Data:**
For training data, we generate long contexts by augmenting [SlimPajama](https://huggingface.co/datasets/cerebras/SlimPajama-627B).
**Progressive Training Details:**
| | 65K | 262K | 524k | 1048k |
|------------------------|-----------|-----------|-----------|-----------|
| Initialize From | LLaMA-3 8B| 65K | 262K | 524k |
| Sequence Length 2^N | 16 | 18 | 19 | 20 |
| RoPE theta | 15.3 M | 207.1 M | 1.06B | 2.80B |
| Batch Size | 1 | 1 | 16 | 16 |
| Gradient Accumulation Steps | 32 | 16 | 1 | 1 |
| Steps | 30 | 24 | 50 | 50 |
| Total Tokens | 62914560 | 100663296 | 419430400 | 838860800 |
| Learning Rate | 2.00E-05 | 2.00E-05 | 2.00E-05 | 2.00E-05 |
| # GPUs | 8 | 32 | 512 | 512 |
| GPU Type | NVIDIA L40S | NVIDIA L40S | NVIDIA L40S | NVIDIA L40S |
| Minutes to Train (Wall)| 202 | 555 | 61 | 87 |
**Quants**:
- [GGUF](https://huggingface.co/crusoeai/Llama-3-8B-Instruct-1048k-GGUF)
- [MLX-4bit](https://huggingface.co/mlx-community/Llama-3-8B-Instruct-1048k-4bit)
## The Gradient AI Team
https://gradient.ai/
Gradient is accelerating AI transformation across industries. Our AI Foundry incorporates your data to deploy autonomous assistants that power critical operations across your business.
## Contact Us
Drop an email to [[email protected]](mailto:[email protected])
## References
[1] Peng, Bowen, et al. "Yarn: Efficient context window extension of large language models." arXiv preprint arXiv:2309.00071 (2023).
[2] Liu, Hao, et al. "World Model on Million-Length Video And Language With RingAttention." arXiv preprint arXiv:2402.08268 (2024).
[3] https://github.com/jzhang38/EasyContext
----
# Base Model
## Model Details
Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.
**Model developers** Meta
**Variations** Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.
**Input** Models input text only.
**Output** Models generate text and code only.
**Model Architecture** Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.
<table>
<tr>
<td>
</td>
<td><strong>Training Data</strong>
</td>
<td><strong>Params</strong>
</td>
<td><strong>Context length</strong>
</td>
<td><strong>GQA</strong>
</td>
<td><strong>Token count</strong>
</td>
<td><strong>Knowledge cutoff</strong>
</td>
</tr>
<tr>
<td rowspan="2" >Llama 3
</td>
<td rowspan="2" >A new mix of publicly available online data.
</td>
<td>8B
</td>
<td>8k
</td>
<td>Yes
</td>
<td rowspan="2" >15T+
</td>
<td>March, 2023
</td>
</tr>
<tr>
<td>70B
</td>
<td>8k
</td>
<td>Yes
</td>
<td>December, 2023
</td>
</tr>
</table>
**Llama 3 family of models**. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.
**Model Release Date** April 18, 2024.
**Status** This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.
**License** A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license)
Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes).
## Intended Use
**Intended Use Cases** Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.
**Out-of-scope** Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English**.
**Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.
## How to use
This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original `llama3` codebase.
### Use with transformers
You can run conversational inference using the Transformers pipeline abstraction, or by leveraging the Auto classes with the `generate()` function. Let's see examples of both.
#### Transformers pipeline
```python
import transformers
import torch
model_id = "meta-llama/Meta-Llama-3-8B-Instruct"
pipeline = transformers.pipeline(
"text-generation",
model=model_id,
model_kwargs={"torch_dtype": torch.bfloat16},
device_map="auto",
)
messages = [
{"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
{"role": "user", "content": "Who are you?"},
]
prompt = pipeline.tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
terminators = [
pipeline.tokenizer.eos_token_id,
pipeline.tokenizer.convert_tokens_to_ids("<|eot_id|>")
]
outputs = pipeline(
prompt,
max_new_tokens=256,
eos_token_id=terminators,
do_sample=True,
temperature=0.6,
top_p=0.9,
)
print(outputs[0]["generated_text"][len(prompt):])
```
#### Transformers AutoModelForCausalLM
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
model_id = "meta-llama/Meta-Llama-3-8B-Instruct"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.bfloat16,
device_map="auto",
)
messages = [
{"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
{"role": "user", "content": "Who are you?"},
]
input_ids = tokenizer.apply_chat_template(
messages,
add_generation_prompt=True,
return_tensors="pt"
).to(model.device)
terminators = [
tokenizer.eos_token_id,
tokenizer.convert_tokens_to_ids("<|eot_id|>")
]
outputs = model.generate(
input_ids,
max_new_tokens=256,
eos_token_id=terminators,
do_sample=True,
temperature=0.6,
top_p=0.9,
)
response = outputs[0][input_ids.shape[-1]:]
print(tokenizer.decode(response, skip_special_tokens=True))
```
### Use with `llama3`
Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3)
To download Original checkpoints, see the example command below leveraging `huggingface-cli`:
```
huggingface-cli download meta-llama/Meta-Llama-3-8B-Instruct --include "original/*" --local-dir Meta-Llama-3-8B-Instruct
```
For Hugging Face support, we recommend using transformers or TGI, but a similar command works.
## Hardware and Software
**Training Factors** We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.
**Carbon Footprint Pretraining utilized a cumulative** 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.
<table>
<tr>
<td>
</td>
<td><strong>Time (GPU hours)</strong>
</td>
<td><strong>Power Consumption (W)</strong>
</td>
<td><strong>Carbon Emitted(tCO2eq)</strong>
</td>
</tr>
<tr>
<td>Llama 3 8B
</td>
<td>1.3M
</td>
<td>700
</td>
<td>390
</td>
</tr>
<tr>
<td>Llama 3 70B
</td>
<td>6.4M
</td>
<td>700
</td>
<td>1900
</td>
</tr>
<tr>
<td>Total
</td>
<td>7.7M
</td>
<td>
</td>
<td>2290
</td>
</tr>
</table>
**CO2 emissions during pre-training**. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.
## Training Data
**Overview** Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.
**Data Freshness** The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.
## Benchmarks
In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md).
### Base pretrained models
<table>
<tr>
<td><strong>Category</strong>
</td>
<td><strong>Benchmark</strong>
</td>
<td><strong>Llama 3 8B</strong>
</td>
<td><strong>Llama2 7B</strong>
</td>
<td><strong>Llama2 13B</strong>
</td>
<td><strong>Llama 3 70B</strong>
</td>
<td><strong>Llama2 70B</strong>
</td>
</tr>
<tr>
<td rowspan="6" >General
</td>
<td>MMLU (5-shot)
</td>
<td>66.6
</td>
<td>45.7
</td>
<td>53.8
</td>
<td>79.5
</td>
<td>69.7
</td>
</tr>
<tr>
<td>AGIEval English (3-5 shot)
</td>
<td>45.9
</td>
<td>28.8
</td>
<td>38.7
</td>
<td>63.0
</td>
<td>54.8
</td>
</tr>
<tr>
<td>CommonSenseQA (7-shot)
</td>
<td>72.6
</td>
<td>57.6
</td>
<td>67.6
</td>
<td>83.8
</td>
<td>78.7
</td>
</tr>
<tr>
<td>Winogrande (5-shot)
</td>
<td>76.1
</td>
<td>73.3
</td>
<td>75.4
</td>
<td>83.1
</td>
<td>81.8
</td>
</tr>
<tr>
<td>BIG-Bench Hard (3-shot, CoT)
</td>
<td>61.1
</td>
<td>38.1
</td>
<td>47.0
</td>
<td>81.3
</td>
<td>65.7
</td>
</tr>
<tr>
<td>ARC-Challenge (25-shot)
</td>
<td>78.6
</td>
<td>53.7
</td>
<td>67.6
</td>
<td>93.0
</td>
<td>85.3
</td>
</tr>
<tr>
<td>Knowledge reasoning
</td>
<td>TriviaQA-Wiki (5-shot)
</td>
<td>78.5
</td>
<td>72.1
</td>
<td>79.6
</td>
<td>89.7
</td>
<td>87.5
</td>
</tr>
<tr>
<td rowspan="4" >Reading comprehension
</td>
<td>SQuAD (1-shot)
</td>
<td>76.4
</td>
<td>72.2
</td>
<td>72.1
</td>
<td>85.6
</td>
<td>82.6
</td>
</tr>
<tr>
<td>QuAC (1-shot, F1)
</td>
<td>44.4
</td>
<td>39.6
</td>
<td>44.9
</td>
<td>51.1
</td>
<td>49.4
</td>
</tr>
<tr>
<td>BoolQ (0-shot)
</td>
<td>75.7
</td>
<td>65.5
</td>
<td>66.9
</td>
<td>79.0
</td>
<td>73.1
</td>
</tr>
<tr>
<td>DROP (3-shot, F1)
</td>
<td>58.4
</td>
<td>37.9
</td>
<td>49.8
</td>
<td>79.7
</td>
<td>70.2
</td>
</tr>
</table>
### Instruction tuned models
<table>
<tr>
<td><strong>Benchmark</strong>
</td>
<td><strong>Llama 3 8B</strong>
</td>
<td><strong>Llama 2 7B</strong>
</td>
<td><strong>Llama 2 13B</strong>
</td>
<td><strong>Llama 3 70B</strong>
</td>
<td><strong>Llama 2 70B</strong>
</td>
</tr>
<tr>
<td>MMLU (5-shot)
</td>
<td>68.4
</td>
<td>34.1
</td>
<td>47.8
</td>
<td>82.0
</td>
<td>52.9
</td>
</tr>
<tr>
<td>GPQA (0-shot)
</td>
<td>34.2
</td>
<td>21.7
</td>
<td>22.3
</td>
<td>39.5
</td>
<td>21.0
</td>
</tr>
<tr>
<td>HumanEval (0-shot)
</td>
<td>62.2
</td>
<td>7.9
</td>
<td>14.0
</td>
<td>81.7
</td>
<td>25.6
</td>
</tr>
<tr>
<td>GSM-8K (8-shot, CoT)
</td>
<td>79.6
</td>
<td>25.7
</td>
<td>77.4
</td>
<td>93.0
</td>
<td>57.5
</td>
</tr>
<tr>
<td>MATH (4-shot, CoT)
</td>
<td>30.0
</td>
<td>3.8
</td>
<td>6.7
</td>
<td>50.4
</td>
<td>11.6
</td>
</tr>
</table>
### Responsibility & Safety
We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.
Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.
Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.
As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started.
#### Llama 3-Instruct
As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.
<span style="text-decoration:underline;">Safety</span>
For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.
<span style="text-decoration:underline;">Refusals</span>
In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.
We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.
#### Responsible release
In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.
Misuse
If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/).
#### Critical risks
<span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)
We have conducted a two fold assessment of the safety of the model in this area:
* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.
* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).
### <span style="text-decoration:underline;">Cyber Security </span>
We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval).
### <span style="text-decoration:underline;">Child Safety</span>
Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.
### Community
Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama).
Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community.
## Ethical Considerations and Limitations
The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.
But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.
Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide)
## Citation instructions
@article{llama3modelcard,
title={Llama 3 Model Card},
author={AI@Meta},
year={2024},
url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md}
}
## Contributors
Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos | {"language": ["en"], "license": "llama3", "tags": ["meta", "llama-3"], "pipeline_tag": "text-generation"} | blockblockblock/Llama-3-8B-Instruct-Gradient-1048k-bpw5.5-exl2 | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"meta",
"llama-3",
"conversational",
"en",
"license:llama3",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:25:08+00:00 |
null | null | {} | sdinger/distil-large-v3-dv | null | [
"region:us"
] | null | 2024-04-30T03:25:27+00:00 |
|
null | null | {"license": "cc-by-nc-sa-4.0"} | chengfeng111/uploada | null | [
"license:cc-by-nc-sa-4.0",
"region:us"
] | null | 2024-04-30T03:26:00+00:00 |
|
null | null | {} | wahyubagus/pegasus-samsum-coba | null | [
"region:us"
] | null | 2024-04-30T03:27:33+00:00 |
|
text-to-image | diffusers |
# LoRA text2image fine-tuning - lerle144/criminal-sketch-lora-v2-2-test
These are LoRA adaption weights for SujinHwang/criminal-sketch-lora-v2-2. The weights were fine-tuned on the SujinHwang/criminal-sketch-Hr dataset. You can find some example images in the following.
| {"license": "creativeml-openrail-m", "tags": ["stable-diffusion", "stable-diffusion-diffusers", "text-to-image", "diffusers", "lora"], "base_model": "SujinHwang/criminal-sketch-lora-v2-2", "inference": true} | lerle144/criminal-sketch-lora-v2-2-test | null | [
"diffusers",
"safetensors",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"lora",
"base_model:SujinHwang/criminal-sketch-lora-v2-2",
"license:creativeml-openrail-m",
"region:us"
] | null | 2024-04-30T03:28:09+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_4-seqsight_16384_512_56M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_4](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_4) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3725
- F1 Score: 0.8399
- Accuracy: 0.84
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5238 | 1.34 | 200 | 0.4992 | 0.7583 | 0.759 |
| 0.4664 | 2.68 | 400 | 0.4759 | 0.7808 | 0.781 |
| 0.4525 | 4.03 | 600 | 0.4760 | 0.7745 | 0.775 |
| 0.4358 | 5.37 | 800 | 0.4844 | 0.7716 | 0.773 |
| 0.4238 | 6.71 | 1000 | 0.4557 | 0.7950 | 0.795 |
| 0.4169 | 8.05 | 1200 | 0.4595 | 0.7882 | 0.789 |
| 0.4108 | 9.4 | 1400 | 0.4539 | 0.7826 | 0.783 |
| 0.4042 | 10.74 | 1600 | 0.4485 | 0.7886 | 0.789 |
| 0.3971 | 12.08 | 1800 | 0.4477 | 0.7939 | 0.794 |
| 0.3919 | 13.42 | 2000 | 0.4684 | 0.7727 | 0.775 |
| 0.3811 | 14.77 | 2200 | 0.4578 | 0.7848 | 0.786 |
| 0.3794 | 16.11 | 2400 | 0.4393 | 0.7960 | 0.796 |
| 0.3731 | 17.45 | 2600 | 0.4471 | 0.7899 | 0.791 |
| 0.3707 | 18.79 | 2800 | 0.4427 | 0.7944 | 0.795 |
| 0.3643 | 20.13 | 3000 | 0.4433 | 0.7981 | 0.799 |
| 0.3582 | 21.48 | 3200 | 0.4291 | 0.8129 | 0.813 |
| 0.3544 | 22.82 | 3400 | 0.4309 | 0.8119 | 0.812 |
| 0.3506 | 24.16 | 3600 | 0.4334 | 0.8057 | 0.806 |
| 0.3464 | 25.5 | 3800 | 0.4255 | 0.8098 | 0.81 |
| 0.3391 | 26.85 | 4000 | 0.4217 | 0.8109 | 0.811 |
| 0.3367 | 28.19 | 4200 | 0.4245 | 0.8219 | 0.822 |
| 0.3253 | 29.53 | 4400 | 0.4299 | 0.8200 | 0.82 |
| 0.3312 | 30.87 | 4600 | 0.4444 | 0.8040 | 0.805 |
| 0.3237 | 32.21 | 4800 | 0.4221 | 0.8169 | 0.817 |
| 0.3184 | 33.56 | 5000 | 0.4365 | 0.8107 | 0.811 |
| 0.3186 | 34.9 | 5200 | 0.4299 | 0.8158 | 0.816 |
| 0.3184 | 36.24 | 5400 | 0.4207 | 0.8220 | 0.822 |
| 0.3095 | 37.58 | 5600 | 0.4174 | 0.8200 | 0.82 |
| 0.3066 | 38.93 | 5800 | 0.4261 | 0.8238 | 0.824 |
| 0.3029 | 40.27 | 6000 | 0.4202 | 0.8300 | 0.83 |
| 0.2993 | 41.61 | 6200 | 0.4267 | 0.8239 | 0.824 |
| 0.2996 | 42.95 | 6400 | 0.4235 | 0.8239 | 0.824 |
| 0.2979 | 44.3 | 6600 | 0.4129 | 0.8270 | 0.827 |
| 0.2883 | 45.64 | 6800 | 0.4233 | 0.8249 | 0.825 |
| 0.2941 | 46.98 | 7000 | 0.4164 | 0.8310 | 0.831 |
| 0.2904 | 48.32 | 7200 | 0.4172 | 0.8340 | 0.834 |
| 0.2881 | 49.66 | 7400 | 0.4151 | 0.8300 | 0.83 |
| 0.2882 | 51.01 | 7600 | 0.4175 | 0.8309 | 0.831 |
| 0.2845 | 52.35 | 7800 | 0.4098 | 0.8300 | 0.83 |
| 0.2805 | 53.69 | 8000 | 0.4201 | 0.8289 | 0.829 |
| 0.2817 | 55.03 | 8200 | 0.4195 | 0.8289 | 0.829 |
| 0.2817 | 56.38 | 8400 | 0.4110 | 0.8330 | 0.833 |
| 0.2754 | 57.72 | 8600 | 0.4155 | 0.8340 | 0.834 |
| 0.2786 | 59.06 | 8800 | 0.4129 | 0.8330 | 0.833 |
| 0.2782 | 60.4 | 9000 | 0.4129 | 0.8350 | 0.835 |
| 0.2743 | 61.74 | 9200 | 0.4120 | 0.8320 | 0.832 |
| 0.2762 | 63.09 | 9400 | 0.4143 | 0.8330 | 0.833 |
| 0.2744 | 64.43 | 9600 | 0.4114 | 0.8350 | 0.835 |
| 0.2723 | 65.77 | 9800 | 0.4126 | 0.8320 | 0.832 |
| 0.2737 | 67.11 | 10000 | 0.4124 | 0.8330 | 0.833 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_4-seqsight_16384_512_56M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_4-seqsight_16384_512_56M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:28:20+00:00 |
reinforcement-learning | stable-baselines3 |
# **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
...
```
| {"library_name": "stable-baselines3", "tags": ["LunarLander-v2", "deep-reinforcement-learning", "reinforcement-learning", "stable-baselines3"], "model-index": [{"name": "PPO", "results": [{"task": {"type": "reinforcement-learning", "name": "reinforcement-learning"}, "dataset": {"name": "LunarLander-v2", "type": "LunarLander-v2"}, "metrics": [{"type": "mean_reward", "value": "257.27 +/- 8.46", "name": "mean_reward", "verified": false}]}]}]} | slepox/ppo-LunarLander-v2 | null | [
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] | null | 2024-04-30T03:31:44+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_4-seqsight_16384_512_56M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_4](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_4) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5689
- F1 Score: 0.8359
- Accuracy: 0.836
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5118 | 1.34 | 200 | 0.4904 | 0.7663 | 0.767 |
| 0.4534 | 2.68 | 400 | 0.4836 | 0.7746 | 0.776 |
| 0.4343 | 4.03 | 600 | 0.4629 | 0.7898 | 0.79 |
| 0.4143 | 5.37 | 800 | 0.4890 | 0.7653 | 0.768 |
| 0.398 | 6.71 | 1000 | 0.4450 | 0.7938 | 0.794 |
| 0.3865 | 8.05 | 1200 | 0.4402 | 0.7938 | 0.794 |
| 0.3717 | 9.4 | 1400 | 0.4323 | 0.8049 | 0.805 |
| 0.359 | 10.74 | 1600 | 0.4238 | 0.8020 | 0.802 |
| 0.3471 | 12.08 | 1800 | 0.4292 | 0.8150 | 0.815 |
| 0.3308 | 13.42 | 2000 | 0.4555 | 0.7784 | 0.781 |
| 0.3153 | 14.77 | 2200 | 0.4453 | 0.8075 | 0.808 |
| 0.3055 | 16.11 | 2400 | 0.4267 | 0.8190 | 0.819 |
| 0.2911 | 17.45 | 2600 | 0.4425 | 0.8106 | 0.811 |
| 0.2843 | 18.79 | 2800 | 0.4238 | 0.8170 | 0.817 |
| 0.272 | 20.13 | 3000 | 0.4293 | 0.8096 | 0.81 |
| 0.2617 | 21.48 | 3200 | 0.4093 | 0.8300 | 0.83 |
| 0.2499 | 22.82 | 3400 | 0.3980 | 0.8380 | 0.838 |
| 0.2402 | 24.16 | 3600 | 0.4325 | 0.8248 | 0.825 |
| 0.2339 | 25.5 | 3800 | 0.4061 | 0.8370 | 0.837 |
| 0.222 | 26.85 | 4000 | 0.4003 | 0.8440 | 0.844 |
| 0.2146 | 28.19 | 4200 | 0.4357 | 0.8228 | 0.823 |
| 0.2035 | 29.53 | 4400 | 0.4240 | 0.8450 | 0.845 |
| 0.2058 | 30.87 | 4600 | 0.4249 | 0.8267 | 0.827 |
| 0.1899 | 32.21 | 4800 | 0.4143 | 0.8410 | 0.841 |
| 0.1859 | 33.56 | 5000 | 0.4138 | 0.8480 | 0.848 |
| 0.1854 | 34.9 | 5200 | 0.4199 | 0.8380 | 0.838 |
| 0.1842 | 36.24 | 5400 | 0.4070 | 0.8530 | 0.853 |
| 0.1698 | 37.58 | 5600 | 0.4226 | 0.8489 | 0.849 |
| 0.1631 | 38.93 | 5800 | 0.4159 | 0.8500 | 0.85 |
| 0.1642 | 40.27 | 6000 | 0.4137 | 0.8649 | 0.865 |
| 0.1584 | 41.61 | 6200 | 0.4399 | 0.8520 | 0.852 |
| 0.1547 | 42.95 | 6400 | 0.4342 | 0.8640 | 0.864 |
| 0.1542 | 44.3 | 6600 | 0.4220 | 0.8719 | 0.872 |
| 0.1443 | 45.64 | 6800 | 0.4383 | 0.8669 | 0.867 |
| 0.1461 | 46.98 | 7000 | 0.4257 | 0.8690 | 0.869 |
| 0.1398 | 48.32 | 7200 | 0.4355 | 0.8709 | 0.871 |
| 0.141 | 49.66 | 7400 | 0.4439 | 0.8739 | 0.874 |
| 0.1356 | 51.01 | 7600 | 0.4510 | 0.8670 | 0.867 |
| 0.1329 | 52.35 | 7800 | 0.4407 | 0.8759 | 0.876 |
| 0.1282 | 53.69 | 8000 | 0.4516 | 0.8670 | 0.867 |
| 0.1294 | 55.03 | 8200 | 0.4458 | 0.8670 | 0.867 |
| 0.1325 | 56.38 | 8400 | 0.4366 | 0.8759 | 0.876 |
| 0.122 | 57.72 | 8600 | 0.4581 | 0.8729 | 0.873 |
| 0.1205 | 59.06 | 8800 | 0.4654 | 0.8730 | 0.873 |
| 0.1247 | 60.4 | 9000 | 0.4460 | 0.8779 | 0.878 |
| 0.119 | 61.74 | 9200 | 0.4640 | 0.8789 | 0.879 |
| 0.12 | 63.09 | 9400 | 0.4597 | 0.8769 | 0.877 |
| 0.1188 | 64.43 | 9600 | 0.4605 | 0.8729 | 0.873 |
| 0.1136 | 65.77 | 9800 | 0.4631 | 0.8749 | 0.875 |
| 0.119 | 67.11 | 10000 | 0.4627 | 0.8749 | 0.875 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_4-seqsight_16384_512_56M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_4-seqsight_16384_512_56M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:32:48+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_3-seqsight_16384_512_56M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_3](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_3) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5605
- F1 Score: 0.7155
- Accuracy: 0.717
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.6323 | 0.93 | 200 | 0.5864 | 0.6761 | 0.676 |
| 0.604 | 1.87 | 400 | 0.5925 | 0.6680 | 0.67 |
| 0.5959 | 2.8 | 600 | 0.5700 | 0.7064 | 0.707 |
| 0.5892 | 3.74 | 800 | 0.5729 | 0.6996 | 0.7 |
| 0.5847 | 4.67 | 1000 | 0.5634 | 0.7001 | 0.7 |
| 0.5826 | 5.61 | 1200 | 0.5606 | 0.7140 | 0.714 |
| 0.5771 | 6.54 | 1400 | 0.5586 | 0.7109 | 0.711 |
| 0.5755 | 7.48 | 1600 | 0.5539 | 0.7148 | 0.715 |
| 0.5727 | 8.41 | 1800 | 0.5514 | 0.7101 | 0.71 |
| 0.5769 | 9.35 | 2000 | 0.5489 | 0.7214 | 0.722 |
| 0.5681 | 10.28 | 2200 | 0.5664 | 0.6899 | 0.692 |
| 0.5665 | 11.21 | 2400 | 0.5469 | 0.7127 | 0.713 |
| 0.5661 | 12.15 | 2600 | 0.5394 | 0.7127 | 0.714 |
| 0.5619 | 13.08 | 2800 | 0.5405 | 0.7081 | 0.71 |
| 0.5612 | 14.02 | 3000 | 0.5426 | 0.7149 | 0.715 |
| 0.5605 | 14.95 | 3200 | 0.5382 | 0.7158 | 0.717 |
| 0.5558 | 15.89 | 3400 | 0.5402 | 0.7063 | 0.707 |
| 0.5557 | 16.82 | 3600 | 0.5389 | 0.7125 | 0.713 |
| 0.5537 | 17.76 | 3800 | 0.5384 | 0.7211 | 0.722 |
| 0.5543 | 18.69 | 4000 | 0.5476 | 0.7028 | 0.703 |
| 0.5522 | 19.63 | 4200 | 0.5372 | 0.7106 | 0.712 |
| 0.5554 | 20.56 | 4400 | 0.5377 | 0.7120 | 0.713 |
| 0.5501 | 21.5 | 4600 | 0.5500 | 0.6938 | 0.695 |
| 0.5465 | 22.43 | 4800 | 0.5411 | 0.7117 | 0.712 |
| 0.5501 | 23.36 | 5000 | 0.5484 | 0.7025 | 0.703 |
| 0.5476 | 24.3 | 5200 | 0.5544 | 0.6915 | 0.693 |
| 0.5426 | 25.23 | 5400 | 0.5394 | 0.7140 | 0.715 |
| 0.5484 | 26.17 | 5600 | 0.5405 | 0.7128 | 0.713 |
| 0.5429 | 27.1 | 5800 | 0.5361 | 0.7202 | 0.721 |
| 0.5409 | 28.04 | 6000 | 0.5402 | 0.7081 | 0.708 |
| 0.5435 | 28.97 | 6200 | 0.5476 | 0.7098 | 0.71 |
| 0.5431 | 29.91 | 6400 | 0.5422 | 0.716 | 0.716 |
| 0.5435 | 30.84 | 6600 | 0.5392 | 0.7130 | 0.713 |
| 0.5416 | 31.78 | 6800 | 0.5427 | 0.7110 | 0.711 |
| 0.5363 | 32.71 | 7000 | 0.5398 | 0.7061 | 0.706 |
| 0.5432 | 33.64 | 7200 | 0.5391 | 0.7121 | 0.712 |
| 0.5344 | 34.58 | 7400 | 0.5443 | 0.7059 | 0.706 |
| 0.5425 | 35.51 | 7600 | 0.5419 | 0.7089 | 0.709 |
| 0.5392 | 36.45 | 7800 | 0.5415 | 0.7131 | 0.713 |
| 0.5392 | 37.38 | 8000 | 0.5414 | 0.7110 | 0.711 |
| 0.5331 | 38.32 | 8200 | 0.5383 | 0.7029 | 0.703 |
| 0.5366 | 39.25 | 8400 | 0.5389 | 0.7091 | 0.709 |
| 0.5357 | 40.19 | 8600 | 0.5399 | 0.7121 | 0.712 |
| 0.5365 | 41.12 | 8800 | 0.5386 | 0.7131 | 0.713 |
| 0.5374 | 42.06 | 9000 | 0.5400 | 0.7071 | 0.707 |
| 0.5353 | 42.99 | 9200 | 0.5410 | 0.7140 | 0.714 |
| 0.5337 | 43.93 | 9400 | 0.5424 | 0.7080 | 0.708 |
| 0.5329 | 44.86 | 9600 | 0.5400 | 0.7151 | 0.715 |
| 0.535 | 45.79 | 9800 | 0.5399 | 0.7131 | 0.713 |
| 0.531 | 46.73 | 10000 | 0.5406 | 0.7141 | 0.714 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_3-seqsight_16384_512_56M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_3-seqsight_16384_512_56M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:33:18+00:00 |
question-answering | transformers | {} | stefandi/test_qa_model | null | [
"transformers",
"tensorboard",
"safetensors",
"distilbert",
"question-answering",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:33:27+00:00 |
|
text2text-generation | transformers |
<!-- 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. -->
# pegasus-samsum
This model is a fine-tuned version of [google/pegasus-cnn_dailymail](https://huggingface.co/google/pegasus-cnn_dailymail) on the samsum 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: 5e-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 16
- 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
- num_epochs: 1
### Training results
### Framework versions
- Transformers 4.36.0
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.15.0
| {"tags": ["generated_from_trainer"], "datasets": ["samsum"], "base_model": "google/pegasus-cnn_dailymail", "model-index": [{"name": "pegasus-samsum", "results": []}]} | wahyubagus/pegasus-samsum | null | [
"transformers",
"safetensors",
"pegasus",
"text2text-generation",
"generated_from_trainer",
"dataset:samsum",
"base_model:google/pegasus-cnn_dailymail",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:33:48+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_3-seqsight_16384_512_56M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_3](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_3) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5387
- F1 Score: 0.7233
- Accuracy: 0.724
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.6236 | 0.93 | 200 | 0.5710 | 0.7063 | 0.707 |
| 0.5923 | 1.87 | 400 | 0.5855 | 0.6685 | 0.672 |
| 0.5851 | 2.8 | 600 | 0.5611 | 0.7061 | 0.706 |
| 0.577 | 3.74 | 800 | 0.5694 | 0.6833 | 0.685 |
| 0.5707 | 4.67 | 1000 | 0.5513 | 0.7090 | 0.709 |
| 0.568 | 5.61 | 1200 | 0.5441 | 0.7132 | 0.714 |
| 0.5629 | 6.54 | 1400 | 0.5598 | 0.7004 | 0.702 |
| 0.5596 | 7.48 | 1600 | 0.5431 | 0.7118 | 0.712 |
| 0.5558 | 8.41 | 1800 | 0.5456 | 0.7026 | 0.703 |
| 0.5589 | 9.35 | 2000 | 0.5411 | 0.7138 | 0.714 |
| 0.5505 | 10.28 | 2200 | 0.5477 | 0.7016 | 0.702 |
| 0.5471 | 11.21 | 2400 | 0.5433 | 0.7011 | 0.701 |
| 0.5463 | 12.15 | 2600 | 0.5308 | 0.7130 | 0.714 |
| 0.5413 | 13.08 | 2800 | 0.5371 | 0.7186 | 0.72 |
| 0.5396 | 14.02 | 3000 | 0.5463 | 0.6998 | 0.7 |
| 0.5387 | 14.95 | 3200 | 0.5381 | 0.7030 | 0.703 |
| 0.5322 | 15.89 | 3400 | 0.5368 | 0.7019 | 0.702 |
| 0.5333 | 16.82 | 3600 | 0.5417 | 0.6990 | 0.699 |
| 0.529 | 17.76 | 3800 | 0.5406 | 0.7101 | 0.71 |
| 0.5288 | 18.69 | 4000 | 0.5563 | 0.6789 | 0.682 |
| 0.5273 | 19.63 | 4200 | 0.5372 | 0.7140 | 0.714 |
| 0.5271 | 20.56 | 4400 | 0.5376 | 0.7111 | 0.711 |
| 0.522 | 21.5 | 4600 | 0.5691 | 0.6759 | 0.681 |
| 0.519 | 22.43 | 4800 | 0.5440 | 0.7121 | 0.712 |
| 0.5196 | 23.36 | 5000 | 0.5611 | 0.7010 | 0.702 |
| 0.5177 | 24.3 | 5200 | 0.5646 | 0.6848 | 0.687 |
| 0.5113 | 25.23 | 5400 | 0.5443 | 0.7149 | 0.715 |
| 0.5159 | 26.17 | 5600 | 0.5434 | 0.7190 | 0.719 |
| 0.5106 | 27.1 | 5800 | 0.5486 | 0.7047 | 0.705 |
| 0.5086 | 28.04 | 6000 | 0.5510 | 0.7035 | 0.704 |
| 0.5082 | 28.97 | 6200 | 0.5570 | 0.7027 | 0.703 |
| 0.5071 | 29.91 | 6400 | 0.5487 | 0.7091 | 0.709 |
| 0.5079 | 30.84 | 6600 | 0.5429 | 0.7161 | 0.716 |
| 0.5054 | 31.78 | 6800 | 0.5573 | 0.6883 | 0.69 |
| 0.5008 | 32.71 | 7000 | 0.5547 | 0.7055 | 0.706 |
| 0.5048 | 33.64 | 7200 | 0.5499 | 0.7008 | 0.701 |
| 0.497 | 34.58 | 7400 | 0.5622 | 0.6937 | 0.695 |
| 0.5046 | 35.51 | 7600 | 0.5654 | 0.6911 | 0.693 |
| 0.4988 | 36.45 | 7800 | 0.5605 | 0.7058 | 0.706 |
| 0.4993 | 37.38 | 8000 | 0.5611 | 0.7011 | 0.702 |
| 0.4931 | 38.32 | 8200 | 0.5586 | 0.7087 | 0.709 |
| 0.4953 | 39.25 | 8400 | 0.5592 | 0.7127 | 0.713 |
| 0.4936 | 40.19 | 8600 | 0.5568 | 0.7137 | 0.714 |
| 0.4966 | 41.12 | 8800 | 0.5558 | 0.7045 | 0.705 |
| 0.4942 | 42.06 | 9000 | 0.5563 | 0.7128 | 0.713 |
| 0.4936 | 42.99 | 9200 | 0.5620 | 0.6906 | 0.692 |
| 0.4918 | 43.93 | 9400 | 0.5653 | 0.6979 | 0.699 |
| 0.4927 | 44.86 | 9600 | 0.5602 | 0.7014 | 0.702 |
| 0.4935 | 45.79 | 9800 | 0.5582 | 0.7066 | 0.707 |
| 0.4906 | 46.73 | 10000 | 0.5603 | 0.7045 | 0.705 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_3-seqsight_16384_512_56M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_3-seqsight_16384_512_56M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:33:49+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_3-seqsight_16384_512_56M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_3](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_3) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5441
- F1 Score: 0.7119
- Accuracy: 0.713
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.6185 | 0.93 | 200 | 0.5652 | 0.7063 | 0.708 |
| 0.5871 | 1.87 | 400 | 0.5634 | 0.7019 | 0.702 |
| 0.5766 | 2.8 | 600 | 0.5567 | 0.7061 | 0.706 |
| 0.5676 | 3.74 | 800 | 0.5587 | 0.6976 | 0.698 |
| 0.5598 | 4.67 | 1000 | 0.5492 | 0.7011 | 0.701 |
| 0.5548 | 5.61 | 1200 | 0.5416 | 0.7143 | 0.716 |
| 0.5498 | 6.54 | 1400 | 0.5592 | 0.6912 | 0.693 |
| 0.5431 | 7.48 | 1600 | 0.5408 | 0.7229 | 0.723 |
| 0.5397 | 8.41 | 1800 | 0.5419 | 0.7090 | 0.709 |
| 0.5364 | 9.35 | 2000 | 0.5440 | 0.6970 | 0.697 |
| 0.5273 | 10.28 | 2200 | 0.5480 | 0.7039 | 0.704 |
| 0.5224 | 11.21 | 2400 | 0.5513 | 0.7128 | 0.713 |
| 0.5167 | 12.15 | 2600 | 0.5444 | 0.7199 | 0.723 |
| 0.5132 | 13.08 | 2800 | 0.5652 | 0.7145 | 0.715 |
| 0.5075 | 14.02 | 3000 | 0.5483 | 0.7041 | 0.704 |
| 0.5046 | 14.95 | 3200 | 0.5586 | 0.7131 | 0.713 |
| 0.4952 | 15.89 | 3400 | 0.5573 | 0.7011 | 0.701 |
| 0.4919 | 16.82 | 3600 | 0.5788 | 0.6936 | 0.695 |
| 0.4875 | 17.76 | 3800 | 0.5561 | 0.6969 | 0.698 |
| 0.4838 | 18.69 | 4000 | 0.5798 | 0.6883 | 0.69 |
| 0.4806 | 19.63 | 4200 | 0.5578 | 0.7051 | 0.705 |
| 0.4764 | 20.56 | 4400 | 0.5675 | 0.7081 | 0.709 |
| 0.4683 | 21.5 | 4600 | 0.5992 | 0.6835 | 0.687 |
| 0.4647 | 22.43 | 4800 | 0.5799 | 0.7050 | 0.705 |
| 0.4609 | 23.36 | 5000 | 0.5904 | 0.7087 | 0.709 |
| 0.4572 | 24.3 | 5200 | 0.5954 | 0.7047 | 0.706 |
| 0.4496 | 25.23 | 5400 | 0.5849 | 0.7038 | 0.704 |
| 0.4502 | 26.17 | 5600 | 0.5736 | 0.7141 | 0.714 |
| 0.4434 | 27.1 | 5800 | 0.5954 | 0.6944 | 0.695 |
| 0.4376 | 28.04 | 6000 | 0.5963 | 0.6976 | 0.698 |
| 0.4355 | 28.97 | 6200 | 0.6120 | 0.6868 | 0.688 |
| 0.4309 | 29.91 | 6400 | 0.6022 | 0.6938 | 0.694 |
| 0.4302 | 30.84 | 6600 | 0.5848 | 0.7031 | 0.703 |
| 0.4281 | 31.78 | 6800 | 0.6133 | 0.6937 | 0.695 |
| 0.4214 | 32.71 | 7000 | 0.6280 | 0.6909 | 0.692 |
| 0.4215 | 33.64 | 7200 | 0.6059 | 0.6870 | 0.688 |
| 0.4151 | 34.58 | 7400 | 0.6255 | 0.6909 | 0.692 |
| 0.4161 | 35.51 | 7600 | 0.6389 | 0.6873 | 0.689 |
| 0.4104 | 36.45 | 7800 | 0.6465 | 0.6951 | 0.696 |
| 0.4092 | 37.38 | 8000 | 0.6402 | 0.6936 | 0.695 |
| 0.4025 | 38.32 | 8200 | 0.6498 | 0.6830 | 0.685 |
| 0.4001 | 39.25 | 8400 | 0.6416 | 0.6984 | 0.699 |
| 0.4027 | 40.19 | 8600 | 0.6404 | 0.6931 | 0.694 |
| 0.4 | 41.12 | 8800 | 0.6448 | 0.6933 | 0.694 |
| 0.4038 | 42.06 | 9000 | 0.6392 | 0.6889 | 0.69 |
| 0.3969 | 42.99 | 9200 | 0.6414 | 0.6877 | 0.689 |
| 0.3935 | 43.93 | 9400 | 0.6509 | 0.6918 | 0.693 |
| 0.3928 | 44.86 | 9600 | 0.6474 | 0.6900 | 0.691 |
| 0.3935 | 45.79 | 9800 | 0.6462 | 0.6899 | 0.691 |
| 0.3905 | 46.73 | 10000 | 0.6503 | 0.6908 | 0.692 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_3-seqsight_16384_512_56M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_3-seqsight_16384_512_56M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:33:50+00:00 |
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
| {"library_name": "transformers", "tags": []} | justinphan3110/zephyr_short_circuit_init | null | [
"transformers",
"safetensors",
"mistral",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:33:58+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_2-seqsight_16384_512_56M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_2](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_2) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4574
- F1 Score: 0.7978
- Accuracy: 0.798
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5758 | 1.34 | 200 | 0.5394 | 0.7409 | 0.741 |
| 0.5362 | 2.68 | 400 | 0.5386 | 0.7293 | 0.731 |
| 0.5277 | 4.03 | 600 | 0.5242 | 0.7370 | 0.738 |
| 0.5193 | 5.37 | 800 | 0.5173 | 0.7416 | 0.742 |
| 0.5152 | 6.71 | 1000 | 0.5210 | 0.7445 | 0.745 |
| 0.5067 | 8.05 | 1200 | 0.5105 | 0.7486 | 0.749 |
| 0.5021 | 9.4 | 1400 | 0.5034 | 0.7518 | 0.752 |
| 0.5014 | 10.74 | 1600 | 0.5014 | 0.7510 | 0.751 |
| 0.4999 | 12.08 | 1800 | 0.5263 | 0.7504 | 0.752 |
| 0.4979 | 13.42 | 2000 | 0.5061 | 0.7548 | 0.755 |
| 0.4952 | 14.77 | 2200 | 0.4996 | 0.7464 | 0.747 |
| 0.4886 | 16.11 | 2400 | 0.5050 | 0.7449 | 0.745 |
| 0.4934 | 17.45 | 2600 | 0.4985 | 0.7454 | 0.746 |
| 0.4893 | 18.79 | 2800 | 0.4966 | 0.7538 | 0.754 |
| 0.481 | 20.13 | 3000 | 0.5019 | 0.7520 | 0.752 |
| 0.4816 | 21.48 | 3200 | 0.5102 | 0.7586 | 0.759 |
| 0.4875 | 22.82 | 3400 | 0.4969 | 0.7590 | 0.759 |
| 0.4797 | 24.16 | 3600 | 0.4938 | 0.7580 | 0.758 |
| 0.4758 | 25.5 | 3800 | 0.4975 | 0.7580 | 0.758 |
| 0.4788 | 26.85 | 4000 | 0.4969 | 0.7480 | 0.748 |
| 0.4758 | 28.19 | 4200 | 0.4974 | 0.7580 | 0.758 |
| 0.4748 | 29.53 | 4400 | 0.5026 | 0.7559 | 0.756 |
| 0.4728 | 30.87 | 4600 | 0.5035 | 0.7518 | 0.752 |
| 0.474 | 32.21 | 4800 | 0.4971 | 0.7529 | 0.753 |
| 0.4697 | 33.56 | 5000 | 0.5047 | 0.7595 | 0.76 |
| 0.4724 | 34.9 | 5200 | 0.4998 | 0.7538 | 0.754 |
| 0.4648 | 36.24 | 5400 | 0.4934 | 0.7539 | 0.754 |
| 0.4711 | 37.58 | 5600 | 0.4952 | 0.7560 | 0.756 |
| 0.4661 | 38.93 | 5800 | 0.4928 | 0.7475 | 0.748 |
| 0.4659 | 40.27 | 6000 | 0.4908 | 0.7497 | 0.75 |
| 0.4657 | 41.61 | 6200 | 0.4927 | 0.7600 | 0.76 |
| 0.4627 | 42.95 | 6400 | 0.5006 | 0.7519 | 0.752 |
| 0.4607 | 44.3 | 6600 | 0.4966 | 0.7540 | 0.754 |
| 0.4648 | 45.64 | 6800 | 0.5016 | 0.7585 | 0.759 |
| 0.4606 | 46.98 | 7000 | 0.4971 | 0.7600 | 0.76 |
| 0.4605 | 48.32 | 7200 | 0.4979 | 0.7609 | 0.761 |
| 0.4603 | 49.66 | 7400 | 0.4936 | 0.7590 | 0.759 |
| 0.46 | 51.01 | 7600 | 0.4965 | 0.7599 | 0.76 |
| 0.4559 | 52.35 | 7800 | 0.4951 | 0.756 | 0.756 |
| 0.46 | 53.69 | 8000 | 0.4933 | 0.7620 | 0.762 |
| 0.4578 | 55.03 | 8200 | 0.4962 | 0.7540 | 0.754 |
| 0.4563 | 56.38 | 8400 | 0.4986 | 0.7549 | 0.755 |
| 0.458 | 57.72 | 8600 | 0.4937 | 0.7550 | 0.755 |
| 0.4536 | 59.06 | 8800 | 0.4952 | 0.7560 | 0.756 |
| 0.4526 | 60.4 | 9000 | 0.5005 | 0.7538 | 0.754 |
| 0.4562 | 61.74 | 9200 | 0.4950 | 0.7600 | 0.76 |
| 0.4546 | 63.09 | 9400 | 0.4965 | 0.7500 | 0.75 |
| 0.454 | 64.43 | 9600 | 0.4944 | 0.7590 | 0.759 |
| 0.4565 | 65.77 | 9800 | 0.4960 | 0.7540 | 0.754 |
| 0.4511 | 67.11 | 10000 | 0.4952 | 0.7560 | 0.756 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_2-seqsight_16384_512_56M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_2-seqsight_16384_512_56M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:34:16+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_2-seqsight_16384_512_56M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_2](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_2) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4601
- F1 Score: 0.78
- Accuracy: 0.78
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5639 | 1.34 | 200 | 0.5200 | 0.7439 | 0.744 |
| 0.5214 | 2.68 | 400 | 0.5351 | 0.7267 | 0.729 |
| 0.5113 | 4.03 | 600 | 0.5134 | 0.7594 | 0.76 |
| 0.5023 | 5.37 | 800 | 0.5203 | 0.7512 | 0.752 |
| 0.4965 | 6.71 | 1000 | 0.5217 | 0.7561 | 0.757 |
| 0.4886 | 8.05 | 1200 | 0.4998 | 0.7480 | 0.748 |
| 0.4837 | 9.4 | 1400 | 0.5085 | 0.7558 | 0.756 |
| 0.4827 | 10.74 | 1600 | 0.5012 | 0.7559 | 0.756 |
| 0.4782 | 12.08 | 1800 | 0.5032 | 0.7599 | 0.76 |
| 0.4762 | 13.42 | 2000 | 0.5037 | 0.7478 | 0.748 |
| 0.4719 | 14.77 | 2200 | 0.4966 | 0.7423 | 0.744 |
| 0.4644 | 16.11 | 2400 | 0.5055 | 0.7569 | 0.757 |
| 0.4671 | 17.45 | 2600 | 0.4976 | 0.7500 | 0.751 |
| 0.4625 | 18.79 | 2800 | 0.4917 | 0.7620 | 0.762 |
| 0.4517 | 20.13 | 3000 | 0.5015 | 0.7658 | 0.766 |
| 0.4515 | 21.48 | 3200 | 0.5118 | 0.7531 | 0.754 |
| 0.4554 | 22.82 | 3400 | 0.4954 | 0.7579 | 0.758 |
| 0.4432 | 24.16 | 3600 | 0.4895 | 0.7630 | 0.763 |
| 0.4388 | 25.5 | 3800 | 0.5074 | 0.7534 | 0.754 |
| 0.4404 | 26.85 | 4000 | 0.4984 | 0.7630 | 0.763 |
| 0.436 | 28.19 | 4200 | 0.5095 | 0.7587 | 0.76 |
| 0.4339 | 29.53 | 4400 | 0.5092 | 0.7635 | 0.764 |
| 0.4294 | 30.87 | 4600 | 0.4974 | 0.7649 | 0.765 |
| 0.4298 | 32.21 | 4800 | 0.5048 | 0.7569 | 0.757 |
| 0.4245 | 33.56 | 5000 | 0.5115 | 0.7593 | 0.76 |
| 0.4259 | 34.9 | 5200 | 0.5033 | 0.7588 | 0.759 |
| 0.4147 | 36.24 | 5400 | 0.4942 | 0.7589 | 0.759 |
| 0.4188 | 37.58 | 5600 | 0.4994 | 0.7610 | 0.761 |
| 0.4143 | 38.93 | 5800 | 0.4967 | 0.7509 | 0.751 |
| 0.4142 | 40.27 | 6000 | 0.4928 | 0.7536 | 0.754 |
| 0.4087 | 41.61 | 6200 | 0.4946 | 0.7526 | 0.753 |
| 0.4065 | 42.95 | 6400 | 0.5137 | 0.7619 | 0.762 |
| 0.4018 | 44.3 | 6600 | 0.5101 | 0.7550 | 0.755 |
| 0.4057 | 45.64 | 6800 | 0.5173 | 0.7593 | 0.76 |
| 0.4012 | 46.98 | 7000 | 0.5121 | 0.7649 | 0.765 |
| 0.4018 | 48.32 | 7200 | 0.5068 | 0.7590 | 0.759 |
| 0.3956 | 49.66 | 7400 | 0.5081 | 0.7560 | 0.756 |
| 0.4004 | 51.01 | 7600 | 0.5031 | 0.7590 | 0.759 |
| 0.3944 | 52.35 | 7800 | 0.5039 | 0.7570 | 0.757 |
| 0.3957 | 53.69 | 8000 | 0.5015 | 0.7560 | 0.756 |
| 0.3927 | 55.03 | 8200 | 0.5092 | 0.7619 | 0.762 |
| 0.3919 | 56.38 | 8400 | 0.5111 | 0.7590 | 0.759 |
| 0.3898 | 57.72 | 8600 | 0.5086 | 0.7570 | 0.757 |
| 0.3893 | 59.06 | 8800 | 0.5101 | 0.7590 | 0.759 |
| 0.3881 | 60.4 | 9000 | 0.5141 | 0.7570 | 0.757 |
| 0.3892 | 61.74 | 9200 | 0.5090 | 0.752 | 0.752 |
| 0.3859 | 63.09 | 9400 | 0.5115 | 0.7560 | 0.756 |
| 0.3858 | 64.43 | 9600 | 0.5090 | 0.7540 | 0.754 |
| 0.3873 | 65.77 | 9800 | 0.5122 | 0.7560 | 0.756 |
| 0.3813 | 67.11 | 10000 | 0.5116 | 0.7560 | 0.756 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_2-seqsight_16384_512_56M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_2-seqsight_16384_512_56M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:34:33+00:00 |
null | peft |
<!-- 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. -->
# GUE_tf_2-seqsight_16384_512_56M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_tf_2](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_tf_2) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4468
- F1 Score: 0.7817
- Accuracy: 0.782
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5546 | 1.34 | 200 | 0.5131 | 0.7472 | 0.748 |
| 0.5106 | 2.68 | 400 | 0.5252 | 0.7407 | 0.742 |
| 0.5022 | 4.03 | 600 | 0.5040 | 0.7490 | 0.749 |
| 0.4902 | 5.37 | 800 | 0.5119 | 0.7447 | 0.745 |
| 0.4809 | 6.71 | 1000 | 0.5143 | 0.7554 | 0.756 |
| 0.4709 | 8.05 | 1200 | 0.5031 | 0.7540 | 0.754 |
| 0.461 | 9.4 | 1400 | 0.5245 | 0.7354 | 0.736 |
| 0.4542 | 10.74 | 1600 | 0.5103 | 0.7478 | 0.748 |
| 0.4447 | 12.08 | 1800 | 0.5174 | 0.7474 | 0.748 |
| 0.4365 | 13.42 | 2000 | 0.5325 | 0.7521 | 0.753 |
| 0.4246 | 14.77 | 2200 | 0.5126 | 0.7440 | 0.744 |
| 0.4128 | 16.11 | 2400 | 0.5431 | 0.7427 | 0.744 |
| 0.4095 | 17.45 | 2600 | 0.5326 | 0.7347 | 0.735 |
| 0.3975 | 18.79 | 2800 | 0.5421 | 0.7414 | 0.742 |
| 0.3792 | 20.13 | 3000 | 0.5704 | 0.7472 | 0.748 |
| 0.3764 | 21.48 | 3200 | 0.5958 | 0.7329 | 0.735 |
| 0.3754 | 22.82 | 3400 | 0.5755 | 0.7459 | 0.746 |
| 0.3585 | 24.16 | 3600 | 0.5903 | 0.7319 | 0.732 |
| 0.349 | 25.5 | 3800 | 0.6227 | 0.7287 | 0.73 |
| 0.3415 | 26.85 | 4000 | 0.6051 | 0.7379 | 0.738 |
| 0.337 | 28.19 | 4200 | 0.6326 | 0.7416 | 0.743 |
| 0.3275 | 29.53 | 4400 | 0.6562 | 0.7359 | 0.737 |
| 0.322 | 30.87 | 4600 | 0.6220 | 0.7368 | 0.737 |
| 0.3164 | 32.21 | 4800 | 0.6639 | 0.7327 | 0.733 |
| 0.3067 | 33.56 | 5000 | 0.6726 | 0.7305 | 0.731 |
| 0.3023 | 34.9 | 5200 | 0.6646 | 0.7325 | 0.733 |
| 0.2901 | 36.24 | 5400 | 0.6579 | 0.7330 | 0.733 |
| 0.2917 | 37.58 | 5600 | 0.6760 | 0.7327 | 0.733 |
| 0.2862 | 38.93 | 5800 | 0.6572 | 0.7300 | 0.73 |
| 0.2765 | 40.27 | 6000 | 0.6905 | 0.7360 | 0.736 |
| 0.2739 | 41.61 | 6200 | 0.6925 | 0.7230 | 0.723 |
| 0.2687 | 42.95 | 6400 | 0.7251 | 0.7172 | 0.718 |
| 0.2605 | 44.3 | 6600 | 0.7733 | 0.7153 | 0.716 |
| 0.2616 | 45.64 | 6800 | 0.7723 | 0.7130 | 0.714 |
| 0.259 | 46.98 | 7000 | 0.7563 | 0.7192 | 0.72 |
| 0.2511 | 48.32 | 7200 | 0.7477 | 0.7235 | 0.724 |
| 0.2504 | 49.66 | 7400 | 0.7455 | 0.7171 | 0.718 |
| 0.2496 | 51.01 | 7600 | 0.7318 | 0.7249 | 0.725 |
| 0.2401 | 52.35 | 7800 | 0.7606 | 0.7237 | 0.724 |
| 0.2404 | 53.69 | 8000 | 0.7715 | 0.7158 | 0.716 |
| 0.2325 | 55.03 | 8200 | 0.7862 | 0.7144 | 0.715 |
| 0.2325 | 56.38 | 8400 | 0.8022 | 0.7162 | 0.717 |
| 0.2333 | 57.72 | 8600 | 0.7831 | 0.7177 | 0.718 |
| 0.2296 | 59.06 | 8800 | 0.7887 | 0.7125 | 0.713 |
| 0.2255 | 60.4 | 9000 | 0.7916 | 0.7096 | 0.71 |
| 0.2271 | 61.74 | 9200 | 0.8099 | 0.7102 | 0.711 |
| 0.2206 | 63.09 | 9400 | 0.7974 | 0.7097 | 0.71 |
| 0.2226 | 64.43 | 9600 | 0.8002 | 0.7138 | 0.714 |
| 0.224 | 65.77 | 9800 | 0.8024 | 0.7086 | 0.709 |
| 0.2124 | 67.11 | 10000 | 0.8061 | 0.7126 | 0.713 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_tf_2-seqsight_16384_512_56M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_tf_2-seqsight_16384_512_56M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:34:35+00:00 |
null | peft |
<!-- 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. -->
# tiny-llama-20p
This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T](https://huggingface.co/TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T) on the HuggingFaceH4/ultrachat_200k dataset.
It achieves the following results on the evaluation set:
- Loss: nan
## 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: 0.0002
- train_batch_size: 14
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 4
- total_train_batch_size: 224
- total_eval_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.0275 | 0.99 | 96 | nan |
### Framework versions
- PEFT 0.7.1
- Transformers 4.39.0.dev0
- Pytorch 2.2.2+cu121
- Datasets 2.14.6
- Tokenizers 0.15.2 | {"license": "apache-2.0", "library_name": "peft", "tags": ["alignment-handbook", "trl", "sft", "generated_from_trainer"], "datasets": ["HuggingFaceH4/ultrachat_200k"], "base_model": "TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T", "model-index": [{"name": "tiny-llama-20p", "results": []}]} | terry69/tiny-llama-20p | null | [
"peft",
"tensorboard",
"safetensors",
"llama",
"alignment-handbook",
"trl",
"sft",
"generated_from_trainer",
"dataset:HuggingFaceH4/ultrachat_200k",
"base_model:TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T",
"license:apache-2.0",
"region:us"
] | null | 2024-04-30T03:35:25+00:00 |
text-classification | transformers | ## TextAttack Model Card
This `albert` model was fine-tuned using TextAttack. The model was fine-tuned
for 3 epochs with a batch size of 8,
a maximum sequence length of 512, and an initial learning rate of 3e-05.
Since this was a classification task, the model was trained with a cross-entropy loss function.
The best score the model achieved on this task was 0.9503333333333334, as measured by the
eval set accuracy, found after 3 epochs.
For more information, check out [TextAttack on Github](https://github.com/QData/TextAttack). | {"language": ["zh"], "license": "apache-2.0", "metrics": ["accuracy"], "pipeline_tag": "text-classification"} | WangA/albert-base-finetuned-jd | null | [
"transformers",
"safetensors",
"albert",
"text-classification",
"zh",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:35:51+00:00 |
text-to-image | diffusers |
# SDXL LoRA DreamBooth - kuei1026/3d-icon-sdxl-dora-rank-64
<Gallery />
## Model description
### These are kuei1026/3d-icon-sdxl-dora-rank-64 LoRA adaption weights for stabilityai/stable-diffusion-xl-base-1.0.
## Download model
### Use it with UIs such as AUTOMATIC1111, Comfy UI, SD.Next, Invoke
- **LoRA**: download **[`3d-icon-sdxl-dora-rank-64.safetensors` here 💾](/kuei1026/3d-icon-sdxl-dora-rank-64/blob/main/3d-icon-sdxl-dora-rank-64.safetensors)**.
- Place it on your `models/Lora` folder.
- On AUTOMATIC1111, load the LoRA by adding `<lora:3d-icon-sdxl-dora-rank-64:1>` to your prompt. On ComfyUI just [load it as a regular LoRA](https://comfyanonymous.github.io/ComfyUI_examples/lora/).
- *Embeddings*: download **[`3d-icon-sdxl-dora-rank-64_emb.safetensors` here 💾](/kuei1026/3d-icon-sdxl-dora-rank-64/blob/main/3d-icon-sdxl-dora-rank-64_emb.safetensors)**.
- Place it on it on your `embeddings` folder
- Use it by adding `3d-icon-sdxl-dora-rank-64_emb` to your prompt. For example, `3d icon in the style of 3d-icon-sdxl-dora-rank-64_emb`
(you need both the LoRA and the embeddings as they were trained together for this LoRA)
## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
```py
from diffusers import AutoPipelineForText2Image
import torch
from huggingface_hub import hf_hub_download
from safetensors.torch import load_file
pipeline = AutoPipelineForText2Image.from_pretrained('stabilityai/stable-diffusion-xl-base-1.0', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('kuei1026/3d-icon-sdxl-dora-rank-64', weight_name='pytorch_lora_weights.safetensors')
embedding_path = hf_hub_download(repo_id='kuei1026/3d-icon-sdxl-dora-rank-64', filename='3d-icon-sdxl-dora-rank-64_emb.safetensors', repo_type="model")
state_dict = load_file(embedding_path)
pipeline.load_textual_inversion(state_dict["clip_l"], token=["<s0>", "<s1>"], text_encoder=pipeline.text_encoder, tokenizer=pipeline.tokenizer)
pipeline.load_textual_inversion(state_dict["clip_g"], token=["<s0>", "<s1>"], text_encoder=pipeline.text_encoder_2, tokenizer=pipeline.tokenizer_2)
image = pipeline('a <s0><s1> icon of an astronaut riding a horse, in the style of <s0><s1>').images[0]
```
For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
## Trigger words
To trigger image generation of trained concept(or concepts) replace each concept identifier in you prompt with the new inserted tokens:
to trigger concept `TOK` → use `<s0><s1>` in your prompt
## Details
All [Files & versions](/kuei1026/3d-icon-sdxl-dora-rank-64/tree/main).
The weights were trained using [🧨 diffusers Advanced Dreambooth Training Script](https://github.com/huggingface/diffusers/blob/main/examples/advanced_diffusion_training/train_dreambooth_lora_sdxl_advanced.py).
LoRA for the text encoder was enabled. False.
Pivotal tuning was enabled: True.
Special VAE used for training: stabilityai/sdxl-vae.
| {"license": "openrail++", "tags": ["stable-diffusion-xl", "stable-diffusion-xl-diffusers", "text-to-image", "diffusers", "lora", "template:sd-lora"], "widget": [{"text": "a <s0><s1> icon of an astronaut riding a horse, in the style of <s0><s1>", "output": {"url": "image_0.png"}}, {"text": "a <s0><s1> icon of an astronaut riding a horse, in the style of <s0><s1>", "output": {"url": "image_1.png"}}, {"text": "a <s0><s1> icon of an astronaut riding a horse, in the style of <s0><s1>", "output": {"url": "image_2.png"}}, {"text": "a <s0><s1> icon of an astronaut riding a horse, in the style of <s0><s1>", "output": {"url": "image_3.png"}}], "base_model": "stabilityai/stable-diffusion-xl-base-1.0", "instance_prompt": "3d icon in the style of <s0><s1>"} | kuei1026/3d-icon-sdxl-dora-rank-64 | null | [
"diffusers",
"stable-diffusion-xl",
"stable-diffusion-xl-diffusers",
"text-to-image",
"lora",
"template:sd-lora",
"base_model:stabilityai/stable-diffusion-xl-base-1.0",
"license:openrail++",
"region:us"
] | null | 2024-04-30T03:38:05+00:00 |
text-generation | transformers | {} | Tristan/pythia-410m-deduped-es-smoothed-no-peft | null | [
"transformers",
"tensorboard",
"safetensors",
"gpt_neox",
"text-generation",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:38:19+00:00 |
|
null | peft |
<!-- 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. -->
# GUE_virus_covid-seqsight_16384_512_56M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_virus_covid](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_virus_covid) dataset.
It achieves the following results on the evaluation set:
- Loss: 1.5162
- F1 Score: 0.4371
- Accuracy: 0.4371
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 2.182 | 0.35 | 200 | 2.1775 | 0.0841 | 0.1412 |
| 2.1726 | 0.7 | 400 | 2.1660 | 0.1066 | 0.1451 |
| 2.1564 | 1.05 | 600 | 2.1569 | 0.1407 | 0.1735 |
| 2.1424 | 1.4 | 800 | 2.1483 | 0.1252 | 0.1600 |
| 2.1238 | 1.75 | 1000 | 2.1317 | 0.1522 | 0.1821 |
| 2.101 | 2.09 | 1200 | 2.0960 | 0.1861 | 0.2114 |
| 2.0618 | 2.44 | 1400 | 2.0383 | 0.2182 | 0.2329 |
| 2.0219 | 2.79 | 1600 | 2.0039 | 0.2315 | 0.2439 |
| 1.9864 | 3.14 | 1800 | 1.9692 | 0.2525 | 0.2630 |
| 1.9665 | 3.49 | 2000 | 1.9435 | 0.2633 | 0.2745 |
| 1.9498 | 3.84 | 2200 | 1.9265 | 0.2657 | 0.2715 |
| 1.9271 | 4.19 | 2400 | 1.9110 | 0.2771 | 0.2814 |
| 1.9063 | 4.54 | 2600 | 1.8943 | 0.2886 | 0.2899 |
| 1.8983 | 4.89 | 2800 | 1.8805 | 0.2899 | 0.2954 |
| 1.8824 | 5.24 | 3000 | 1.8511 | 0.3031 | 0.3089 |
| 1.8662 | 5.58 | 3200 | 1.8352 | 0.3065 | 0.3087 |
| 1.849 | 5.93 | 3400 | 1.8327 | 0.2973 | 0.2949 |
| 1.832 | 6.28 | 3600 | 1.8001 | 0.3253 | 0.3236 |
| 1.8094 | 6.63 | 3800 | 1.7848 | 0.3223 | 0.3225 |
| 1.8079 | 6.98 | 4000 | 1.7734 | 0.3324 | 0.3346 |
| 1.7856 | 7.33 | 4200 | 1.7566 | 0.3366 | 0.3368 |
| 1.7747 | 7.68 | 4400 | 1.7540 | 0.3356 | 0.3341 |
| 1.766 | 8.03 | 4600 | 1.7129 | 0.3660 | 0.3657 |
| 1.7401 | 8.38 | 4800 | 1.7202 | 0.3530 | 0.3475 |
| 1.7468 | 8.73 | 5000 | 1.7120 | 0.3533 | 0.3549 |
| 1.7303 | 9.08 | 5200 | 1.6838 | 0.3702 | 0.3662 |
| 1.7167 | 9.42 | 5400 | 1.6934 | 0.3541 | 0.3574 |
| 1.7073 | 9.77 | 5600 | 1.6669 | 0.3789 | 0.3782 |
| 1.7026 | 10.12 | 5800 | 1.6605 | 0.3737 | 0.3697 |
| 1.6895 | 10.47 | 6000 | 1.6411 | 0.3844 | 0.3876 |
| 1.6799 | 10.82 | 6200 | 1.6305 | 0.3825 | 0.3825 |
| 1.6634 | 11.17 | 6400 | 1.6153 | 0.3933 | 0.3935 |
| 1.6594 | 11.52 | 6600 | 1.6044 | 0.3991 | 0.3988 |
| 1.6549 | 11.87 | 6800 | 1.5947 | 0.3991 | 0.4001 |
| 1.643 | 12.22 | 7000 | 1.5902 | 0.4036 | 0.4077 |
| 1.6278 | 12.57 | 7200 | 1.5767 | 0.4151 | 0.4149 |
| 1.6357 | 12.91 | 7400 | 1.5633 | 0.4126 | 0.4174 |
| 1.6289 | 13.26 | 7600 | 1.5600 | 0.4130 | 0.4163 |
| 1.6177 | 13.61 | 7800 | 1.5497 | 0.4240 | 0.4279 |
| 1.6138 | 13.96 | 8000 | 1.5474 | 0.4227 | 0.4230 |
| 1.5984 | 14.31 | 8200 | 1.5354 | 0.4307 | 0.4310 |
| 1.6059 | 14.66 | 8400 | 1.5368 | 0.4261 | 0.4248 |
| 1.5988 | 15.01 | 8600 | 1.5352 | 0.4250 | 0.4265 |
| 1.6015 | 15.36 | 8800 | 1.5282 | 0.4288 | 0.4286 |
| 1.5812 | 15.71 | 9000 | 1.5261 | 0.4312 | 0.4352 |
| 1.587 | 16.06 | 9200 | 1.5240 | 0.4297 | 0.4309 |
| 1.5875 | 16.4 | 9400 | 1.5220 | 0.4323 | 0.4331 |
| 1.588 | 16.75 | 9600 | 1.5218 | 0.4282 | 0.4301 |
| 1.5881 | 17.1 | 9800 | 1.5218 | 0.4281 | 0.4301 |
| 1.5856 | 17.45 | 10000 | 1.5215 | 0.4293 | 0.4308 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_virus_covid-seqsight_16384_512_56M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_virus_covid-seqsight_16384_512_56M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:40:25+00:00 |
null | null | {} | leoho0722/Llama-2-7b-chat-hf-GGUF | null | [
"gguf",
"region:us"
] | null | 2024-04-30T03:40:30+00:00 |
|
text-generation | transformers | <a href="https://www.gradient.ai" target="_blank"><img src="https://cdn-uploads.huggingface.co/production/uploads/655bb613e8a8971e89944f3e/TSa3V8YpoVagnTYgxiLaO.png" width="200"/></a>
# Llama-3 8B Gradient Instruct 1048k
Gradient incorporates your data to deploy autonomous assistants that power critical operations across your business. If you're looking to build custom AI models or agents, email us a message [email protected].
For more info see our [End-to-end development service for custom LLMs and AI systems](https://gradient.ai/development-lab)
This model extends LLama-3 8B's context length from 8k to > 1040K, developed by Gradient, sponsored by compute from [Crusoe Energy](https://huggingface.co/crusoeai). It demonstrates that SOTA LLMs can learn to operate on long context with minimal training by appropriately adjusting RoPE theta. We trained on 830M tokens for this stage, and 1.4B tokens total for all stages, which is < 0.01% of Llama-3's original pre-training data.
**Approach:**
- [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) as the base
- NTK-aware interpolation [1] to initialize an optimal schedule for RoPE theta, followed by empirical RoPE theta optimization
- Progressive training on increasing context lengths, similar to [Large World Model](https://huggingface.co/LargeWorldModel) [2] (See details below)
**Infra:**
We build on top of the EasyContext Blockwise RingAttention library [3] to scalably and efficiently train on contexts up to 1048k tokens on [Crusoe Energy](https://huggingface.co/crusoeai) high performance L40S cluster.
Notably, we layered parallelism on top of Ring Attention with a custom network topology to better leverage large GPU clusters in the face of network bottlenecks from passing many KV blocks between devices. This gave us a 33x speedup in model training (compare 524k and 1048k to 65k and 262k in the table below).
**Data:**
For training data, we generate long contexts by augmenting [SlimPajama](https://huggingface.co/datasets/cerebras/SlimPajama-627B).
**Progressive Training Details:**
| | 65K | 262K | 524k | 1048k |
|------------------------|-----------|-----------|-----------|-----------|
| Initialize From | LLaMA-3 8B| 65K | 262K | 524k |
| Sequence Length 2^N | 16 | 18 | 19 | 20 |
| RoPE theta | 15.3 M | 207.1 M | 1.06B | 2.80B |
| Batch Size | 1 | 1 | 16 | 16 |
| Gradient Accumulation Steps | 32 | 16 | 1 | 1 |
| Steps | 30 | 24 | 50 | 50 |
| Total Tokens | 62914560 | 100663296 | 419430400 | 838860800 |
| Learning Rate | 2.00E-05 | 2.00E-05 | 2.00E-05 | 2.00E-05 |
| # GPUs | 8 | 32 | 512 | 512 |
| GPU Type | NVIDIA L40S | NVIDIA L40S | NVIDIA L40S | NVIDIA L40S |
| Minutes to Train (Wall)| 202 | 555 | 61 | 87 |
**Quants**:
- [GGUF](https://huggingface.co/crusoeai/Llama-3-8B-Instruct-1048k-GGUF)
- [MLX-4bit](https://huggingface.co/mlx-community/Llama-3-8B-Instruct-1048k-4bit)
## The Gradient AI Team
https://gradient.ai/
Gradient is accelerating AI transformation across industries. Our AI Foundry incorporates your data to deploy autonomous assistants that power critical operations across your business.
## Contact Us
Drop an email to [[email protected]](mailto:[email protected])
## References
[1] Peng, Bowen, et al. "Yarn: Efficient context window extension of large language models." arXiv preprint arXiv:2309.00071 (2023).
[2] Liu, Hao, et al. "World Model on Million-Length Video And Language With RingAttention." arXiv preprint arXiv:2402.08268 (2024).
[3] https://github.com/jzhang38/EasyContext
----
# Base Model
## Model Details
Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.
**Model developers** Meta
**Variations** Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.
**Input** Models input text only.
**Output** Models generate text and code only.
**Model Architecture** Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.
<table>
<tr>
<td>
</td>
<td><strong>Training Data</strong>
</td>
<td><strong>Params</strong>
</td>
<td><strong>Context length</strong>
</td>
<td><strong>GQA</strong>
</td>
<td><strong>Token count</strong>
</td>
<td><strong>Knowledge cutoff</strong>
</td>
</tr>
<tr>
<td rowspan="2" >Llama 3
</td>
<td rowspan="2" >A new mix of publicly available online data.
</td>
<td>8B
</td>
<td>8k
</td>
<td>Yes
</td>
<td rowspan="2" >15T+
</td>
<td>March, 2023
</td>
</tr>
<tr>
<td>70B
</td>
<td>8k
</td>
<td>Yes
</td>
<td>December, 2023
</td>
</tr>
</table>
**Llama 3 family of models**. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.
**Model Release Date** April 18, 2024.
**Status** This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.
**License** A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license)
Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes).
## Intended Use
**Intended Use Cases** Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.
**Out-of-scope** Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English**.
**Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.
## How to use
This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original `llama3` codebase.
### Use with transformers
You can run conversational inference using the Transformers pipeline abstraction, or by leveraging the Auto classes with the `generate()` function. Let's see examples of both.
#### Transformers pipeline
```python
import transformers
import torch
model_id = "meta-llama/Meta-Llama-3-8B-Instruct"
pipeline = transformers.pipeline(
"text-generation",
model=model_id,
model_kwargs={"torch_dtype": torch.bfloat16},
device_map="auto",
)
messages = [
{"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
{"role": "user", "content": "Who are you?"},
]
prompt = pipeline.tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
terminators = [
pipeline.tokenizer.eos_token_id,
pipeline.tokenizer.convert_tokens_to_ids("<|eot_id|>")
]
outputs = pipeline(
prompt,
max_new_tokens=256,
eos_token_id=terminators,
do_sample=True,
temperature=0.6,
top_p=0.9,
)
print(outputs[0]["generated_text"][len(prompt):])
```
#### Transformers AutoModelForCausalLM
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
model_id = "meta-llama/Meta-Llama-3-8B-Instruct"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.bfloat16,
device_map="auto",
)
messages = [
{"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
{"role": "user", "content": "Who are you?"},
]
input_ids = tokenizer.apply_chat_template(
messages,
add_generation_prompt=True,
return_tensors="pt"
).to(model.device)
terminators = [
tokenizer.eos_token_id,
tokenizer.convert_tokens_to_ids("<|eot_id|>")
]
outputs = model.generate(
input_ids,
max_new_tokens=256,
eos_token_id=terminators,
do_sample=True,
temperature=0.6,
top_p=0.9,
)
response = outputs[0][input_ids.shape[-1]:]
print(tokenizer.decode(response, skip_special_tokens=True))
```
### Use with `llama3`
Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3)
To download Original checkpoints, see the example command below leveraging `huggingface-cli`:
```
huggingface-cli download meta-llama/Meta-Llama-3-8B-Instruct --include "original/*" --local-dir Meta-Llama-3-8B-Instruct
```
For Hugging Face support, we recommend using transformers or TGI, but a similar command works.
## Hardware and Software
**Training Factors** We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.
**Carbon Footprint Pretraining utilized a cumulative** 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.
<table>
<tr>
<td>
</td>
<td><strong>Time (GPU hours)</strong>
</td>
<td><strong>Power Consumption (W)</strong>
</td>
<td><strong>Carbon Emitted(tCO2eq)</strong>
</td>
</tr>
<tr>
<td>Llama 3 8B
</td>
<td>1.3M
</td>
<td>700
</td>
<td>390
</td>
</tr>
<tr>
<td>Llama 3 70B
</td>
<td>6.4M
</td>
<td>700
</td>
<td>1900
</td>
</tr>
<tr>
<td>Total
</td>
<td>7.7M
</td>
<td>
</td>
<td>2290
</td>
</tr>
</table>
**CO2 emissions during pre-training**. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.
## Training Data
**Overview** Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.
**Data Freshness** The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.
## Benchmarks
In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md).
### Base pretrained models
<table>
<tr>
<td><strong>Category</strong>
</td>
<td><strong>Benchmark</strong>
</td>
<td><strong>Llama 3 8B</strong>
</td>
<td><strong>Llama2 7B</strong>
</td>
<td><strong>Llama2 13B</strong>
</td>
<td><strong>Llama 3 70B</strong>
</td>
<td><strong>Llama2 70B</strong>
</td>
</tr>
<tr>
<td rowspan="6" >General
</td>
<td>MMLU (5-shot)
</td>
<td>66.6
</td>
<td>45.7
</td>
<td>53.8
</td>
<td>79.5
</td>
<td>69.7
</td>
</tr>
<tr>
<td>AGIEval English (3-5 shot)
</td>
<td>45.9
</td>
<td>28.8
</td>
<td>38.7
</td>
<td>63.0
</td>
<td>54.8
</td>
</tr>
<tr>
<td>CommonSenseQA (7-shot)
</td>
<td>72.6
</td>
<td>57.6
</td>
<td>67.6
</td>
<td>83.8
</td>
<td>78.7
</td>
</tr>
<tr>
<td>Winogrande (5-shot)
</td>
<td>76.1
</td>
<td>73.3
</td>
<td>75.4
</td>
<td>83.1
</td>
<td>81.8
</td>
</tr>
<tr>
<td>BIG-Bench Hard (3-shot, CoT)
</td>
<td>61.1
</td>
<td>38.1
</td>
<td>47.0
</td>
<td>81.3
</td>
<td>65.7
</td>
</tr>
<tr>
<td>ARC-Challenge (25-shot)
</td>
<td>78.6
</td>
<td>53.7
</td>
<td>67.6
</td>
<td>93.0
</td>
<td>85.3
</td>
</tr>
<tr>
<td>Knowledge reasoning
</td>
<td>TriviaQA-Wiki (5-shot)
</td>
<td>78.5
</td>
<td>72.1
</td>
<td>79.6
</td>
<td>89.7
</td>
<td>87.5
</td>
</tr>
<tr>
<td rowspan="4" >Reading comprehension
</td>
<td>SQuAD (1-shot)
</td>
<td>76.4
</td>
<td>72.2
</td>
<td>72.1
</td>
<td>85.6
</td>
<td>82.6
</td>
</tr>
<tr>
<td>QuAC (1-shot, F1)
</td>
<td>44.4
</td>
<td>39.6
</td>
<td>44.9
</td>
<td>51.1
</td>
<td>49.4
</td>
</tr>
<tr>
<td>BoolQ (0-shot)
</td>
<td>75.7
</td>
<td>65.5
</td>
<td>66.9
</td>
<td>79.0
</td>
<td>73.1
</td>
</tr>
<tr>
<td>DROP (3-shot, F1)
</td>
<td>58.4
</td>
<td>37.9
</td>
<td>49.8
</td>
<td>79.7
</td>
<td>70.2
</td>
</tr>
</table>
### Instruction tuned models
<table>
<tr>
<td><strong>Benchmark</strong>
</td>
<td><strong>Llama 3 8B</strong>
</td>
<td><strong>Llama 2 7B</strong>
</td>
<td><strong>Llama 2 13B</strong>
</td>
<td><strong>Llama 3 70B</strong>
</td>
<td><strong>Llama 2 70B</strong>
</td>
</tr>
<tr>
<td>MMLU (5-shot)
</td>
<td>68.4
</td>
<td>34.1
</td>
<td>47.8
</td>
<td>82.0
</td>
<td>52.9
</td>
</tr>
<tr>
<td>GPQA (0-shot)
</td>
<td>34.2
</td>
<td>21.7
</td>
<td>22.3
</td>
<td>39.5
</td>
<td>21.0
</td>
</tr>
<tr>
<td>HumanEval (0-shot)
</td>
<td>62.2
</td>
<td>7.9
</td>
<td>14.0
</td>
<td>81.7
</td>
<td>25.6
</td>
</tr>
<tr>
<td>GSM-8K (8-shot, CoT)
</td>
<td>79.6
</td>
<td>25.7
</td>
<td>77.4
</td>
<td>93.0
</td>
<td>57.5
</td>
</tr>
<tr>
<td>MATH (4-shot, CoT)
</td>
<td>30.0
</td>
<td>3.8
</td>
<td>6.7
</td>
<td>50.4
</td>
<td>11.6
</td>
</tr>
</table>
### Responsibility & Safety
We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.
Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.
Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.
As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started.
#### Llama 3-Instruct
As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.
<span style="text-decoration:underline;">Safety</span>
For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.
<span style="text-decoration:underline;">Refusals</span>
In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.
We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.
#### Responsible release
In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.
Misuse
If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/).
#### Critical risks
<span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)
We have conducted a two fold assessment of the safety of the model in this area:
* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.
* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).
### <span style="text-decoration:underline;">Cyber Security </span>
We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval).
### <span style="text-decoration:underline;">Child Safety</span>
Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.
### Community
Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama).
Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community.
## Ethical Considerations and Limitations
The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.
But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.
Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide)
## Citation instructions
@article{llama3modelcard,
title={Llama 3 Model Card},
author={AI@Meta},
year={2024},
url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md}
}
## Contributors
Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos | {"language": ["en"], "license": "llama3", "tags": ["meta", "llama-3"], "pipeline_tag": "text-generation"} | blockblockblock/Llama-3-8B-Instruct-Gradient-1048k-bpw6-exl2 | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"meta",
"llama-3",
"conversational",
"en",
"license:llama3",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"6-bit",
"region:us"
] | null | 2024-04-30T03:40:34+00:00 |
null | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": ["unsloth"]} | trex5790/model_l | null | [
"transformers",
"safetensors",
"unsloth",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:40:40+00:00 |
null | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": ["unsloth"]} | trex5790/model_l2 | null | [
"transformers",
"safetensors",
"unsloth",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:42:22+00:00 |
null | null | {"license": "apache-2.0"} | abovzv/sdxl_segmentation_controlnet_ade20k | null | [
"license:apache-2.0",
"region:us"
] | null | 2024-04-30T03:42:57+00:00 |
|
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | KaifengGGG/Llama-2-7b-spider | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:43:03+00:00 |
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | cohesionet/gemma-1.1-2b-it-test | null | [
"transformers",
"safetensors",
"gemma",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:43:41+00:00 |
text-generation | transformers |
<!-- 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. -->
# meditron-7b-dpo-full-sft-wo-healthsearch_qa
This model is a fine-tuned version of [Minbyul/meditron-7b-wo-healthsearch_qa-sft](https://huggingface.co/Minbyul/meditron-7b-wo-healthsearch_qa-sft) on the HuggingFaceH4/ultrafeedback_binarized dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6911
- Rewards/chosen: 0.0010
- Rewards/rejected: -0.0039
- Rewards/accuracies: 0.6566
- Rewards/margins: 0.0049
- Logps/rejected: -893.7659
- Logps/chosen: -564.5244
- Logits/rejected: -0.7763
- Logits/chosen: -0.8804
## 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-07
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1
### Training results
### Framework versions
- Transformers 4.39.0.dev0
- Pytorch 2.1.2
- Datasets 2.14.6
- Tokenizers 0.15.2
| {"license": "llama2", "tags": ["alignment-handbook", "trl", "dpo", "generated_from_trainer", "trl", "dpo", "generated_from_trainer"], "datasets": ["HuggingFaceH4/ultrafeedback_binarized"], "base_model": "Minbyul/meditron-7b-wo-healthsearch_qa-sft", "model-index": [{"name": "meditron-7b-dpo-full-sft-wo-healthsearch_qa", "results": []}]} | Minbyul/meditron-7b-dpo-full-sft-wo-healthsearch_qa | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"alignment-handbook",
"trl",
"dpo",
"generated_from_trainer",
"dataset:HuggingFaceH4/ultrafeedback_binarized",
"base_model:Minbyul/meditron-7b-wo-healthsearch_qa-sft",
"license:llama2",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:44:10+00:00 |
image-classification | transformers |
# Model Trained Using AutoTrain
- Problem type: Image Classification
## Validation Metrics
loss: 0.2648889720439911
f1_macro: 0.7457675172458867
f1_micro: 0.904404233526801
f1_weighted: 0.9015634064092323
precision_macro: 0.8165440763859227
precision_micro: 0.904404233526801
precision_weighted: 0.9040411595949299
recall_macro: 0.7142884694149322
recall_micro: 0.904404233526801
recall_weighted: 0.904404233526801
accuracy: 0.904404233526801
| {"tags": ["autotrain", "image-classification"], "datasets": ["autotrain-9e6d1-2u0z9/autotrain-data"], "widget": [{"src": "https://huggingface.co/datasets/mishig/sample_images/resolve/main/tiger.jpg", "example_title": "Tiger"}, {"src": "https://huggingface.co/datasets/mishig/sample_images/resolve/main/teapot.jpg", "example_title": "Teapot"}, {"src": "https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg", "example_title": "Palace"}]} | Kushagra07/autotrain-9e6d1-2u0z9 | null | [
"transformers",
"tensorboard",
"safetensors",
"vit",
"image-classification",
"autotrain",
"dataset:autotrain-9e6d1-2u0z9/autotrain-data",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:44:40+00:00 |
null | peft |
<!-- 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. -->
# GUE_virus_covid-seqsight_16384_512_56M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_virus_covid](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_virus_covid) dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1462
- F1 Score: 0.5613
- Accuracy: 0.5659
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 2.1811 | 0.35 | 200 | 2.1751 | 0.0952 | 0.1422 |
| 2.1644 | 0.7 | 400 | 2.1552 | 0.1200 | 0.1561 |
| 2.1297 | 1.05 | 600 | 2.1113 | 0.1730 | 0.2005 |
| 2.0628 | 1.4 | 800 | 1.9971 | 0.2151 | 0.2390 |
| 1.9674 | 1.75 | 1000 | 1.9221 | 0.2697 | 0.2808 |
| 1.9113 | 2.09 | 1200 | 1.8530 | 0.2962 | 0.3059 |
| 1.8395 | 2.44 | 1400 | 1.7629 | 0.3390 | 0.3309 |
| 1.77 | 2.79 | 1600 | 1.6973 | 0.3520 | 0.3606 |
| 1.7051 | 3.14 | 1800 | 1.6203 | 0.3923 | 0.3971 |
| 1.6492 | 3.49 | 2000 | 1.5539 | 0.4003 | 0.4142 |
| 1.6207 | 3.84 | 2200 | 1.5116 | 0.4202 | 0.4278 |
| 1.5801 | 4.19 | 2400 | 1.5038 | 0.4170 | 0.4216 |
| 1.5536 | 4.54 | 2600 | 1.4649 | 0.4304 | 0.4349 |
| 1.5351 | 4.89 | 2800 | 1.4407 | 0.4417 | 0.4428 |
| 1.5127 | 5.24 | 3000 | 1.4229 | 0.4442 | 0.4518 |
| 1.4958 | 5.58 | 3200 | 1.3847 | 0.4661 | 0.4713 |
| 1.4827 | 5.93 | 3400 | 1.3756 | 0.4565 | 0.4689 |
| 1.4527 | 6.28 | 3600 | 1.3477 | 0.4793 | 0.4843 |
| 1.4282 | 6.63 | 3800 | 1.3230 | 0.4912 | 0.4975 |
| 1.4309 | 6.98 | 4000 | 1.3328 | 0.4844 | 0.4938 |
| 1.4063 | 7.33 | 4200 | 1.3018 | 0.5048 | 0.5085 |
| 1.3974 | 7.68 | 4400 | 1.2799 | 0.5123 | 0.5157 |
| 1.3873 | 8.03 | 4600 | 1.2810 | 0.5049 | 0.5064 |
| 1.363 | 8.38 | 4800 | 1.2610 | 0.5171 | 0.5197 |
| 1.3709 | 8.73 | 5000 | 1.2577 | 0.5232 | 0.5265 |
| 1.3491 | 9.08 | 5200 | 1.2477 | 0.5288 | 0.5280 |
| 1.3379 | 9.42 | 5400 | 1.2359 | 0.5185 | 0.5257 |
| 1.3322 | 9.77 | 5600 | 1.2224 | 0.5362 | 0.5413 |
| 1.3228 | 10.12 | 5800 | 1.2131 | 0.5419 | 0.5433 |
| 1.3209 | 10.47 | 6000 | 1.2122 | 0.5430 | 0.5455 |
| 1.3078 | 10.82 | 6200 | 1.2107 | 0.5402 | 0.5392 |
| 1.2893 | 11.17 | 6400 | 1.2002 | 0.5380 | 0.5374 |
| 1.294 | 11.52 | 6600 | 1.1974 | 0.5469 | 0.5470 |
| 1.2944 | 11.87 | 6800 | 1.1875 | 0.5421 | 0.5452 |
| 1.2892 | 12.22 | 7000 | 1.1848 | 0.5489 | 0.5533 |
| 1.2641 | 12.57 | 7200 | 1.1789 | 0.5568 | 0.5563 |
| 1.2773 | 12.91 | 7400 | 1.1743 | 0.5550 | 0.5537 |
| 1.2658 | 13.26 | 7600 | 1.1704 | 0.5543 | 0.5540 |
| 1.2623 | 13.61 | 7800 | 1.1723 | 0.5591 | 0.5597 |
| 1.2626 | 13.96 | 8000 | 1.1675 | 0.5598 | 0.5612 |
| 1.2483 | 14.31 | 8200 | 1.1677 | 0.5546 | 0.5561 |
| 1.2562 | 14.66 | 8400 | 1.1570 | 0.5578 | 0.5599 |
| 1.2497 | 15.01 | 8600 | 1.1554 | 0.5571 | 0.5586 |
| 1.252 | 15.36 | 8800 | 1.1560 | 0.5619 | 0.5622 |
| 1.2315 | 15.71 | 9000 | 1.1552 | 0.5584 | 0.5607 |
| 1.2471 | 16.06 | 9200 | 1.1521 | 0.5587 | 0.5612 |
| 1.2399 | 16.4 | 9400 | 1.1516 | 0.5610 | 0.5640 |
| 1.241 | 16.75 | 9600 | 1.1482 | 0.5627 | 0.5661 |
| 1.2416 | 17.1 | 9800 | 1.1486 | 0.5624 | 0.5652 |
| 1.2399 | 17.45 | 10000 | 1.1485 | 0.5628 | 0.5659 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_virus_covid-seqsight_16384_512_56M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_virus_covid-seqsight_16384_512_56M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:44:46+00:00 |
null | peft |
<!-- 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. -->
# GUE_virus_covid-seqsight_16384_512_56M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_16384_512_56M](https://huggingface.co/mahdibaghbanzadeh/seqsight_16384_512_56M) on the [mahdibaghbanzadeh/GUE_virus_covid](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_virus_covid) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9754
- F1 Score: 0.6278
- Accuracy: 0.6266
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 2.1804 | 0.35 | 200 | 2.1720 | 0.1035 | 0.1394 |
| 2.1553 | 0.7 | 400 | 2.1291 | 0.1636 | 0.1870 |
| 2.0472 | 1.05 | 600 | 1.9548 | 0.2347 | 0.2564 |
| 1.9344 | 1.4 | 800 | 1.8740 | 0.2620 | 0.2885 |
| 1.8076 | 1.75 | 1000 | 1.7210 | 0.3531 | 0.3581 |
| 1.7134 | 2.09 | 1200 | 1.6262 | 0.3819 | 0.3856 |
| 1.6229 | 2.44 | 1400 | 1.5153 | 0.4418 | 0.4363 |
| 1.5368 | 2.79 | 1600 | 1.4230 | 0.4505 | 0.4633 |
| 1.4651 | 3.14 | 1800 | 1.3462 | 0.4944 | 0.4954 |
| 1.4113 | 3.49 | 2000 | 1.3083 | 0.4962 | 0.5 |
| 1.3788 | 3.84 | 2200 | 1.2749 | 0.5128 | 0.5100 |
| 1.3319 | 4.19 | 2400 | 1.2309 | 0.5444 | 0.5326 |
| 1.3146 | 4.54 | 2600 | 1.2101 | 0.5519 | 0.5435 |
| 1.2868 | 4.89 | 2800 | 1.1824 | 0.5587 | 0.5540 |
| 1.2548 | 5.24 | 3000 | 1.1670 | 0.5592 | 0.5557 |
| 1.2455 | 5.58 | 3200 | 1.1601 | 0.5630 | 0.5585 |
| 1.2274 | 5.93 | 3400 | 1.1470 | 0.5521 | 0.5544 |
| 1.2131 | 6.28 | 3600 | 1.1294 | 0.5816 | 0.5715 |
| 1.1833 | 6.63 | 3800 | 1.1076 | 0.5833 | 0.5777 |
| 1.1763 | 6.98 | 4000 | 1.1071 | 0.5811 | 0.5742 |
| 1.1636 | 7.33 | 4200 | 1.0874 | 0.5880 | 0.5853 |
| 1.1414 | 7.68 | 4400 | 1.0691 | 0.5954 | 0.5894 |
| 1.1469 | 8.03 | 4600 | 1.0654 | 0.5883 | 0.5868 |
| 1.1177 | 8.38 | 4800 | 1.0573 | 0.5994 | 0.5948 |
| 1.1268 | 8.73 | 5000 | 1.0438 | 0.5978 | 0.5955 |
| 1.1053 | 9.08 | 5200 | 1.0406 | 0.6044 | 0.5962 |
| 1.0983 | 9.42 | 5400 | 1.0321 | 0.6000 | 0.5972 |
| 1.0932 | 9.77 | 5600 | 1.0275 | 0.6037 | 0.5986 |
| 1.0911 | 10.12 | 5800 | 1.0229 | 0.6063 | 0.6038 |
| 1.0806 | 10.47 | 6000 | 1.0201 | 0.6031 | 0.6000 |
| 1.0766 | 10.82 | 6200 | 1.0163 | 0.6119 | 0.6034 |
| 1.0617 | 11.17 | 6400 | 1.0137 | 0.6090 | 0.6034 |
| 1.0587 | 11.52 | 6600 | 1.0042 | 0.6137 | 0.6093 |
| 1.0662 | 11.87 | 6800 | 1.0059 | 0.6133 | 0.6099 |
| 1.0594 | 12.22 | 7000 | 0.9989 | 0.6139 | 0.6091 |
| 1.0363 | 12.57 | 7200 | 0.9958 | 0.6127 | 0.6094 |
| 1.0497 | 12.91 | 7400 | 0.9955 | 0.6181 | 0.6136 |
| 1.0437 | 13.26 | 7600 | 0.9909 | 0.6176 | 0.6113 |
| 1.0387 | 13.61 | 7800 | 0.9841 | 0.6208 | 0.6158 |
| 1.0367 | 13.96 | 8000 | 0.9807 | 0.6253 | 0.6219 |
| 1.0249 | 14.31 | 8200 | 0.9805 | 0.6241 | 0.6196 |
| 1.0271 | 14.66 | 8400 | 0.9801 | 0.6247 | 0.6191 |
| 1.0315 | 15.01 | 8600 | 0.9761 | 0.6216 | 0.6175 |
| 1.0267 | 15.36 | 8800 | 0.9774 | 0.6240 | 0.6166 |
| 1.0106 | 15.71 | 9000 | 0.9728 | 0.6306 | 0.6278 |
| 1.023 | 16.06 | 9200 | 0.9719 | 0.6262 | 0.6218 |
| 1.0229 | 16.4 | 9400 | 0.9697 | 0.6253 | 0.6236 |
| 1.0158 | 16.75 | 9600 | 0.9703 | 0.6275 | 0.6246 |
| 1.0157 | 17.1 | 9800 | 0.9700 | 0.6272 | 0.6238 |
| 1.0166 | 17.45 | 10000 | 0.9700 | 0.6269 | 0.6230 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_16384_512_56M", "model-index": [{"name": "GUE_virus_covid-seqsight_16384_512_56M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_virus_covid-seqsight_16384_512_56M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_16384_512_56M",
"region:us"
] | null | 2024-04-30T03:45:24+00:00 |
null | null | {} | sda231/led-large-16384-pubmed | null | [
"region:us"
] | null | 2024-04-30T03:45:30+00:00 |
|
null | peft |
<!-- 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. -->
# GUE_prom_prom_300_tata-seqsight_32768_512_30M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_32768_512_30M](https://huggingface.co/mahdibaghbanzadeh/seqsight_32768_512_30M) on the [mahdibaghbanzadeh/GUE_prom_prom_300_tata](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_prom_prom_300_tata) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4768
- F1 Score: 0.7977
- Accuracy: 0.7977
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:------:|:-----:|:---------------:|:--------:|:--------:|
| 0.5988 | 5.13 | 200 | 0.5451 | 0.7335 | 0.7357 |
| 0.4941 | 10.26 | 400 | 0.5144 | 0.7636 | 0.7635 |
| 0.4614 | 15.38 | 600 | 0.4944 | 0.7750 | 0.7749 |
| 0.4439 | 20.51 | 800 | 0.4870 | 0.7848 | 0.7847 |
| 0.4331 | 25.64 | 1000 | 0.4870 | 0.7864 | 0.7863 |
| 0.4206 | 30.77 | 1200 | 0.4789 | 0.7832 | 0.7830 |
| 0.414 | 35.9 | 1400 | 0.4811 | 0.7979 | 0.7977 |
| 0.4044 | 41.03 | 1600 | 0.4792 | 0.8042 | 0.8042 |
| 0.4019 | 46.15 | 1800 | 0.4860 | 0.7897 | 0.7896 |
| 0.3942 | 51.28 | 2000 | 0.4814 | 0.8060 | 0.8059 |
| 0.3901 | 56.41 | 2200 | 0.4869 | 0.8074 | 0.8075 |
| 0.3847 | 61.54 | 2400 | 0.4934 | 0.8027 | 0.8026 |
| 0.378 | 66.67 | 2600 | 0.5069 | 0.7978 | 0.7977 |
| 0.3736 | 71.79 | 2800 | 0.4990 | 0.8060 | 0.8059 |
| 0.3721 | 76.92 | 3000 | 0.5256 | 0.7861 | 0.7863 |
| 0.3681 | 82.05 | 3200 | 0.5077 | 0.7961 | 0.7961 |
| 0.3654 | 87.18 | 3400 | 0.5271 | 0.7877 | 0.7879 |
| 0.3618 | 92.31 | 3600 | 0.5198 | 0.7945 | 0.7945 |
| 0.3564 | 97.44 | 3800 | 0.5158 | 0.8011 | 0.8010 |
| 0.3563 | 102.56 | 4000 | 0.5249 | 0.7897 | 0.7896 |
| 0.3502 | 107.69 | 4200 | 0.5294 | 0.7928 | 0.7928 |
| 0.3502 | 112.82 | 4400 | 0.5256 | 0.7929 | 0.7928 |
| 0.3483 | 117.95 | 4600 | 0.5296 | 0.7945 | 0.7945 |
| 0.3456 | 123.08 | 4800 | 0.5315 | 0.8043 | 0.8042 |
| 0.3415 | 128.21 | 5000 | 0.5252 | 0.8027 | 0.8026 |
| 0.3381 | 133.33 | 5200 | 0.5261 | 0.8076 | 0.8075 |
| 0.3431 | 138.46 | 5400 | 0.5161 | 0.7995 | 0.7993 |
| 0.3352 | 143.59 | 5600 | 0.5347 | 0.7995 | 0.7993 |
| 0.3351 | 148.72 | 5800 | 0.5320 | 0.7995 | 0.7993 |
| 0.3345 | 153.85 | 6000 | 0.5329 | 0.7995 | 0.7993 |
| 0.3292 | 158.97 | 6200 | 0.5435 | 0.7862 | 0.7863 |
| 0.3281 | 164.1 | 6400 | 0.5403 | 0.7994 | 0.7993 |
| 0.3269 | 169.23 | 6600 | 0.5462 | 0.7928 | 0.7928 |
| 0.3247 | 174.36 | 6800 | 0.5509 | 0.7813 | 0.7814 |
| 0.3247 | 179.49 | 7000 | 0.5414 | 0.7929 | 0.7928 |
| 0.3192 | 184.62 | 7200 | 0.5430 | 0.8028 | 0.8026 |
| 0.3234 | 189.74 | 7400 | 0.5531 | 0.7943 | 0.7945 |
| 0.319 | 194.87 | 7600 | 0.5489 | 0.7946 | 0.7945 |
| 0.3173 | 200.0 | 7800 | 0.5478 | 0.7979 | 0.7977 |
| 0.3222 | 205.13 | 8000 | 0.5446 | 0.7912 | 0.7912 |
| 0.3162 | 210.26 | 8200 | 0.5501 | 0.7896 | 0.7896 |
| 0.3161 | 215.38 | 8400 | 0.5491 | 0.7895 | 0.7896 |
| 0.3146 | 220.51 | 8600 | 0.5480 | 0.7978 | 0.7977 |
| 0.3149 | 225.64 | 8800 | 0.5583 | 0.7960 | 0.7961 |
| 0.3149 | 230.77 | 9000 | 0.5558 | 0.7961 | 0.7961 |
| 0.3127 | 235.9 | 9200 | 0.5549 | 0.7961 | 0.7961 |
| 0.3159 | 241.03 | 9400 | 0.5491 | 0.7946 | 0.7945 |
| 0.3138 | 246.15 | 9600 | 0.5539 | 0.7945 | 0.7945 |
| 0.3138 | 251.28 | 9800 | 0.5537 | 0.7945 | 0.7945 |
| 0.3123 | 256.41 | 10000 | 0.5537 | 0.7961 | 0.7961 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_32768_512_30M", "model-index": [{"name": "GUE_prom_prom_300_tata-seqsight_32768_512_30M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_prom_prom_300_tata-seqsight_32768_512_30M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_32768_512_30M",
"region:us"
] | null | 2024-04-30T03:46:03+00:00 |
null | peft |
<!-- 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. -->
# GUE_prom_prom_300_tata-seqsight_32768_512_30M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_32768_512_30M](https://huggingface.co/mahdibaghbanzadeh/seqsight_32768_512_30M) on the [mahdibaghbanzadeh/GUE_prom_prom_300_tata](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_prom_prom_300_tata) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4621
- F1 Score: 0.8043
- Accuracy: 0.8042
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:------:|:-----:|:---------------:|:--------:|:--------:|
| 0.551 | 5.13 | 200 | 0.5087 | 0.7565 | 0.7586 |
| 0.4518 | 10.26 | 400 | 0.4925 | 0.7881 | 0.7879 |
| 0.4222 | 15.38 | 600 | 0.4762 | 0.8059 | 0.8059 |
| 0.3967 | 20.51 | 800 | 0.4751 | 0.8059 | 0.8059 |
| 0.3853 | 25.64 | 1000 | 0.5032 | 0.7845 | 0.7847 |
| 0.364 | 30.77 | 1200 | 0.4943 | 0.8011 | 0.8010 |
| 0.3499 | 35.9 | 1400 | 0.5057 | 0.8011 | 0.8010 |
| 0.332 | 41.03 | 1600 | 0.5050 | 0.8026 | 0.8026 |
| 0.3273 | 46.15 | 1800 | 0.5209 | 0.7946 | 0.7945 |
| 0.3114 | 51.28 | 2000 | 0.5312 | 0.7979 | 0.7977 |
| 0.3004 | 56.41 | 2200 | 0.5695 | 0.7872 | 0.7879 |
| 0.2895 | 61.54 | 2400 | 0.5624 | 0.7957 | 0.7961 |
| 0.277 | 66.67 | 2600 | 0.5815 | 0.7913 | 0.7912 |
| 0.2677 | 71.79 | 2800 | 0.6145 | 0.7926 | 0.7928 |
| 0.263 | 76.92 | 3000 | 0.5911 | 0.7783 | 0.7781 |
| 0.254 | 82.05 | 3200 | 0.6300 | 0.7881 | 0.7879 |
| 0.246 | 87.18 | 3400 | 0.6234 | 0.7846 | 0.7847 |
| 0.2384 | 92.31 | 3600 | 0.6357 | 0.7881 | 0.7879 |
| 0.2302 | 97.44 | 3800 | 0.6524 | 0.7860 | 0.7863 |
| 0.2257 | 102.56 | 4000 | 0.6910 | 0.7814 | 0.7814 |
| 0.2148 | 107.69 | 4200 | 0.6822 | 0.7782 | 0.7781 |
| 0.2143 | 112.82 | 4400 | 0.6871 | 0.7783 | 0.7781 |
| 0.2082 | 117.95 | 4600 | 0.6927 | 0.7881 | 0.7879 |
| 0.2059 | 123.08 | 4800 | 0.7218 | 0.7815 | 0.7814 |
| 0.1956 | 128.21 | 5000 | 0.7267 | 0.7799 | 0.7798 |
| 0.1913 | 133.33 | 5200 | 0.7532 | 0.7864 | 0.7863 |
| 0.1913 | 138.46 | 5400 | 0.7614 | 0.7750 | 0.7749 |
| 0.1842 | 143.59 | 5600 | 0.7697 | 0.7766 | 0.7765 |
| 0.1799 | 148.72 | 5800 | 0.7743 | 0.7734 | 0.7732 |
| 0.1774 | 153.85 | 6000 | 0.7841 | 0.7734 | 0.7732 |
| 0.1738 | 158.97 | 6200 | 0.8049 | 0.7748 | 0.7749 |
| 0.1733 | 164.1 | 6400 | 0.8029 | 0.7767 | 0.7765 |
| 0.1673 | 169.23 | 6600 | 0.7966 | 0.7782 | 0.7781 |
| 0.1642 | 174.36 | 6800 | 0.8250 | 0.7734 | 0.7732 |
| 0.1653 | 179.49 | 7000 | 0.7927 | 0.7750 | 0.7749 |
| 0.1581 | 184.62 | 7200 | 0.8363 | 0.7732 | 0.7732 |
| 0.1597 | 189.74 | 7400 | 0.8298 | 0.7783 | 0.7781 |
| 0.1556 | 194.87 | 7600 | 0.8436 | 0.7783 | 0.7781 |
| 0.1544 | 200.0 | 7800 | 0.8535 | 0.7767 | 0.7765 |
| 0.1553 | 205.13 | 8000 | 0.8489 | 0.7832 | 0.7830 |
| 0.1561 | 210.26 | 8200 | 0.8514 | 0.7750 | 0.7749 |
| 0.1517 | 215.38 | 8400 | 0.8410 | 0.7799 | 0.7798 |
| 0.1486 | 220.51 | 8600 | 0.8643 | 0.7783 | 0.7781 |
| 0.1431 | 225.64 | 8800 | 0.8938 | 0.7718 | 0.7716 |
| 0.1423 | 230.77 | 9000 | 0.8926 | 0.7733 | 0.7732 |
| 0.15 | 235.9 | 9200 | 0.8668 | 0.7783 | 0.7781 |
| 0.1463 | 241.03 | 9400 | 0.8815 | 0.7783 | 0.7781 |
| 0.1449 | 246.15 | 9600 | 0.8718 | 0.7783 | 0.7781 |
| 0.1421 | 251.28 | 9800 | 0.8722 | 0.7767 | 0.7765 |
| 0.1433 | 256.41 | 10000 | 0.8732 | 0.7783 | 0.7781 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_32768_512_30M", "model-index": [{"name": "GUE_prom_prom_300_tata-seqsight_32768_512_30M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_prom_prom_300_tata-seqsight_32768_512_30M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_32768_512_30M",
"region:us"
] | null | 2024-04-30T03:46:08+00:00 |
null | peft |
<!-- 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. -->
# GUE_prom_prom_300_tata-seqsight_32768_512_30M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_32768_512_30M](https://huggingface.co/mahdibaghbanzadeh/seqsight_32768_512_30M) on the [mahdibaghbanzadeh/GUE_prom_prom_300_tata](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_prom_prom_300_tata) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4893
- F1 Score: 0.8092
- Accuracy: 0.8091
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:------:|:-----:|:---------------:|:--------:|:--------:|
| 0.5269 | 5.13 | 200 | 0.4896 | 0.7863 | 0.7863 |
| 0.428 | 10.26 | 400 | 0.5296 | 0.7671 | 0.7684 |
| 0.3818 | 15.38 | 600 | 0.4850 | 0.8027 | 0.8026 |
| 0.3409 | 20.51 | 800 | 0.5115 | 0.8028 | 0.8026 |
| 0.3155 | 25.64 | 1000 | 0.5550 | 0.7975 | 0.7977 |
| 0.2779 | 30.77 | 1200 | 0.5595 | 0.8026 | 0.8026 |
| 0.253 | 35.9 | 1400 | 0.5794 | 0.7909 | 0.7912 |
| 0.2203 | 41.03 | 1600 | 0.6487 | 0.7995 | 0.7993 |
| 0.2103 | 46.15 | 1800 | 0.6713 | 0.7943 | 0.7945 |
| 0.1845 | 51.28 | 2000 | 0.7026 | 0.7994 | 0.7993 |
| 0.1676 | 56.41 | 2200 | 0.7440 | 0.7991 | 0.7993 |
| 0.1569 | 61.54 | 2400 | 0.7793 | 0.7989 | 0.7993 |
| 0.1407 | 66.67 | 2600 | 0.7914 | 0.7979 | 0.7977 |
| 0.1295 | 71.79 | 2800 | 0.8457 | 0.7927 | 0.7928 |
| 0.1234 | 76.92 | 3000 | 0.7828 | 0.8027 | 0.8026 |
| 0.1184 | 82.05 | 3200 | 0.8599 | 0.8028 | 0.8026 |
| 0.1053 | 87.18 | 3400 | 0.9115 | 0.7876 | 0.7879 |
| 0.1038 | 92.31 | 3600 | 0.9341 | 0.7896 | 0.7896 |
| 0.093 | 97.44 | 3800 | 0.9623 | 0.7945 | 0.7945 |
| 0.0918 | 102.56 | 4000 | 1.0186 | 0.7993 | 0.7993 |
| 0.0875 | 107.69 | 4200 | 1.0039 | 0.7828 | 0.7830 |
| 0.0826 | 112.82 | 4400 | 1.0375 | 0.7946 | 0.7945 |
| 0.0772 | 117.95 | 4600 | 1.0244 | 0.7846 | 0.7847 |
| 0.0742 | 123.08 | 4800 | 1.0708 | 0.7994 | 0.7993 |
| 0.0702 | 128.21 | 5000 | 1.0825 | 0.7961 | 0.7961 |
| 0.0679 | 133.33 | 5200 | 1.0522 | 0.7962 | 0.7961 |
| 0.0658 | 138.46 | 5400 | 1.0907 | 0.7979 | 0.7977 |
| 0.0639 | 143.59 | 5600 | 1.0721 | 0.7897 | 0.7896 |
| 0.0576 | 148.72 | 5800 | 1.1193 | 0.7864 | 0.7863 |
| 0.0596 | 153.85 | 6000 | 1.1812 | 0.7962 | 0.7961 |
| 0.0611 | 158.97 | 6200 | 1.0850 | 0.7930 | 0.7928 |
| 0.0575 | 164.1 | 6400 | 1.1514 | 0.7831 | 0.7830 |
| 0.0519 | 169.23 | 6600 | 1.1475 | 0.7913 | 0.7912 |
| 0.05 | 174.36 | 6800 | 1.2358 | 0.7994 | 0.7993 |
| 0.0487 | 179.49 | 7000 | 1.1894 | 0.7962 | 0.7961 |
| 0.0507 | 184.62 | 7200 | 1.2145 | 0.7764 | 0.7765 |
| 0.0495 | 189.74 | 7400 | 1.2251 | 0.7847 | 0.7847 |
| 0.0446 | 194.87 | 7600 | 1.2608 | 0.7879 | 0.7879 |
| 0.0461 | 200.0 | 7800 | 1.2584 | 0.7880 | 0.7879 |
| 0.0457 | 205.13 | 8000 | 1.2233 | 0.7897 | 0.7896 |
| 0.0453 | 210.26 | 8200 | 1.2514 | 0.7946 | 0.7945 |
| 0.0475 | 215.38 | 8400 | 1.2118 | 0.7863 | 0.7863 |
| 0.0434 | 220.51 | 8600 | 1.2464 | 0.7880 | 0.7879 |
| 0.0407 | 225.64 | 8800 | 1.2793 | 0.7946 | 0.7945 |
| 0.0398 | 230.77 | 9000 | 1.3253 | 0.7830 | 0.7830 |
| 0.0415 | 235.9 | 9200 | 1.2875 | 0.7879 | 0.7879 |
| 0.041 | 241.03 | 9400 | 1.2933 | 0.7913 | 0.7912 |
| 0.0407 | 246.15 | 9600 | 1.3033 | 0.7864 | 0.7863 |
| 0.0396 | 251.28 | 9800 | 1.2820 | 0.7929 | 0.7928 |
| 0.0402 | 256.41 | 10000 | 1.2833 | 0.7946 | 0.7945 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_32768_512_30M", "model-index": [{"name": "GUE_prom_prom_300_tata-seqsight_32768_512_30M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_prom_prom_300_tata-seqsight_32768_512_30M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_32768_512_30M",
"region:us"
] | null | 2024-04-30T03:46:38+00:00 |
null | peft |
<!-- 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. -->
# GUE_prom_prom_300_notata-seqsight_32768_512_30M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_32768_512_30M](https://huggingface.co/mahdibaghbanzadeh/seqsight_32768_512_30M) on the [mahdibaghbanzadeh/GUE_prom_prom_300_notata](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_prom_prom_300_notata) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1274
- F1 Score: 0.9510
- Accuracy: 0.9510
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.3546 | 0.6 | 200 | 0.1804 | 0.9274 | 0.9275 |
| 0.1868 | 1.2 | 400 | 0.1557 | 0.9389 | 0.9389 |
| 0.1688 | 1.81 | 600 | 0.1406 | 0.9482 | 0.9482 |
| 0.1517 | 2.41 | 800 | 0.1386 | 0.9487 | 0.9487 |
| 0.1481 | 3.01 | 1000 | 0.1335 | 0.9497 | 0.9497 |
| 0.1431 | 3.61 | 1200 | 0.1276 | 0.9497 | 0.9497 |
| 0.1434 | 4.22 | 1400 | 0.1280 | 0.9493 | 0.9493 |
| 0.1376 | 4.82 | 1600 | 0.1255 | 0.9525 | 0.9525 |
| 0.1353 | 5.42 | 1800 | 0.1248 | 0.9516 | 0.9516 |
| 0.1314 | 6.02 | 2000 | 0.1231 | 0.9521 | 0.9521 |
| 0.1343 | 6.63 | 2200 | 0.1214 | 0.9529 | 0.9529 |
| 0.1269 | 7.23 | 2400 | 0.1219 | 0.9532 | 0.9533 |
| 0.1255 | 7.83 | 2600 | 0.1209 | 0.9540 | 0.9540 |
| 0.1287 | 8.43 | 2800 | 0.1222 | 0.9529 | 0.9529 |
| 0.1316 | 9.04 | 3000 | 0.1251 | 0.9515 | 0.9516 |
| 0.1258 | 9.64 | 3200 | 0.1177 | 0.9551 | 0.9552 |
| 0.1233 | 10.24 | 3400 | 0.1205 | 0.9553 | 0.9553 |
| 0.126 | 10.84 | 3600 | 0.1185 | 0.9546 | 0.9546 |
| 0.1213 | 11.45 | 3800 | 0.1155 | 0.9555 | 0.9555 |
| 0.1212 | 12.05 | 4000 | 0.1157 | 0.9561 | 0.9561 |
| 0.1251 | 12.65 | 4200 | 0.1136 | 0.9568 | 0.9568 |
| 0.1254 | 13.25 | 4400 | 0.1161 | 0.9549 | 0.9550 |
| 0.1205 | 13.86 | 4600 | 0.1151 | 0.9555 | 0.9555 |
| 0.1215 | 14.46 | 4800 | 0.1168 | 0.9546 | 0.9546 |
| 0.1192 | 15.06 | 5000 | 0.1136 | 0.9566 | 0.9567 |
| 0.1214 | 15.66 | 5200 | 0.1130 | 0.9567 | 0.9567 |
| 0.1232 | 16.27 | 5400 | 0.1150 | 0.9559 | 0.9559 |
| 0.1177 | 16.87 | 5600 | 0.1123 | 0.9568 | 0.9568 |
| 0.1155 | 17.47 | 5800 | 0.1121 | 0.9572 | 0.9572 |
| 0.1227 | 18.07 | 6000 | 0.1125 | 0.9568 | 0.9568 |
| 0.1188 | 18.67 | 6200 | 0.1121 | 0.9568 | 0.9568 |
| 0.1201 | 19.28 | 6400 | 0.1134 | 0.9570 | 0.9570 |
| 0.1188 | 19.88 | 6600 | 0.1136 | 0.9568 | 0.9568 |
| 0.1231 | 20.48 | 6800 | 0.1119 | 0.9576 | 0.9576 |
| 0.1138 | 21.08 | 7000 | 0.1124 | 0.9585 | 0.9585 |
| 0.1166 | 21.69 | 7200 | 0.1117 | 0.9567 | 0.9567 |
| 0.1155 | 22.29 | 7400 | 0.1128 | 0.9568 | 0.9568 |
| 0.1205 | 22.89 | 7600 | 0.1107 | 0.9574 | 0.9574 |
| 0.1143 | 23.49 | 7800 | 0.1126 | 0.9582 | 0.9582 |
| 0.1161 | 24.1 | 8000 | 0.1123 | 0.9583 | 0.9584 |
| 0.116 | 24.7 | 8200 | 0.1121 | 0.9583 | 0.9584 |
| 0.1147 | 25.3 | 8400 | 0.1111 | 0.9576 | 0.9576 |
| 0.1158 | 25.9 | 8600 | 0.1116 | 0.9578 | 0.9578 |
| 0.1159 | 26.51 | 8800 | 0.1108 | 0.9578 | 0.9578 |
| 0.1139 | 27.11 | 9000 | 0.1115 | 0.9582 | 0.9582 |
| 0.1179 | 27.71 | 9200 | 0.1115 | 0.9589 | 0.9589 |
| 0.1124 | 28.31 | 9400 | 0.1117 | 0.9583 | 0.9584 |
| 0.1188 | 28.92 | 9600 | 0.1111 | 0.9582 | 0.9582 |
| 0.1119 | 29.52 | 9800 | 0.1112 | 0.9582 | 0.9582 |
| 0.1178 | 30.12 | 10000 | 0.1112 | 0.9580 | 0.9580 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_32768_512_30M", "model-index": [{"name": "GUE_prom_prom_300_notata-seqsight_32768_512_30M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_prom_prom_300_notata-seqsight_32768_512_30M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_32768_512_30M",
"region:us"
] | null | 2024-04-30T03:46:54+00:00 |
null | peft |
<!-- 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. -->
# GUE_prom_prom_300_notata-seqsight_32768_512_30M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_32768_512_30M](https://huggingface.co/mahdibaghbanzadeh/seqsight_32768_512_30M) on the [mahdibaghbanzadeh/GUE_prom_prom_300_notata](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_prom_prom_300_notata) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1229
- F1 Score: 0.9574
- Accuracy: 0.9574
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.299 | 0.6 | 200 | 0.1491 | 0.9435 | 0.9435 |
| 0.1525 | 1.2 | 400 | 0.1315 | 0.9502 | 0.9503 |
| 0.1463 | 1.81 | 600 | 0.1240 | 0.9519 | 0.9520 |
| 0.1322 | 2.41 | 800 | 0.1232 | 0.9529 | 0.9529 |
| 0.1334 | 3.01 | 1000 | 0.1199 | 0.9550 | 0.9550 |
| 0.1276 | 3.61 | 1200 | 0.1198 | 0.9540 | 0.9540 |
| 0.1297 | 4.22 | 1400 | 0.1188 | 0.9538 | 0.9538 |
| 0.1238 | 4.82 | 1600 | 0.1153 | 0.9550 | 0.9550 |
| 0.1226 | 5.42 | 1800 | 0.1172 | 0.9565 | 0.9565 |
| 0.118 | 6.02 | 2000 | 0.1156 | 0.9553 | 0.9553 |
| 0.1212 | 6.63 | 2200 | 0.1134 | 0.9565 | 0.9565 |
| 0.1164 | 7.23 | 2400 | 0.1141 | 0.9584 | 0.9584 |
| 0.1134 | 7.83 | 2600 | 0.1151 | 0.9572 | 0.9572 |
| 0.1163 | 8.43 | 2800 | 0.1129 | 0.9572 | 0.9572 |
| 0.1183 | 9.04 | 3000 | 0.1186 | 0.9546 | 0.9546 |
| 0.1141 | 9.64 | 3200 | 0.1188 | 0.9580 | 0.9580 |
| 0.1104 | 10.24 | 3400 | 0.1136 | 0.9566 | 0.9567 |
| 0.1129 | 10.84 | 3600 | 0.1130 | 0.9591 | 0.9591 |
| 0.1079 | 11.45 | 3800 | 0.1124 | 0.9574 | 0.9574 |
| 0.1091 | 12.05 | 4000 | 0.1128 | 0.9580 | 0.9580 |
| 0.1114 | 12.65 | 4200 | 0.1111 | 0.9584 | 0.9584 |
| 0.112 | 13.25 | 4400 | 0.1129 | 0.9585 | 0.9585 |
| 0.1073 | 13.86 | 4600 | 0.1127 | 0.9591 | 0.9591 |
| 0.1077 | 14.46 | 4800 | 0.1123 | 0.9595 | 0.9595 |
| 0.1068 | 15.06 | 5000 | 0.1105 | 0.9602 | 0.9602 |
| 0.1073 | 15.66 | 5200 | 0.1105 | 0.9600 | 0.9601 |
| 0.1087 | 16.27 | 5400 | 0.1126 | 0.9589 | 0.9589 |
| 0.104 | 16.87 | 5600 | 0.1103 | 0.9608 | 0.9608 |
| 0.1013 | 17.47 | 5800 | 0.1086 | 0.9606 | 0.9606 |
| 0.1081 | 18.07 | 6000 | 0.1090 | 0.9606 | 0.9606 |
| 0.103 | 18.67 | 6200 | 0.1097 | 0.9599 | 0.9599 |
| 0.1051 | 19.28 | 6400 | 0.1133 | 0.9608 | 0.9608 |
| 0.1033 | 19.88 | 6600 | 0.1133 | 0.9597 | 0.9597 |
| 0.107 | 20.48 | 6800 | 0.1093 | 0.9610 | 0.9610 |
| 0.099 | 21.08 | 7000 | 0.1105 | 0.9621 | 0.9621 |
| 0.1015 | 21.69 | 7200 | 0.1128 | 0.9584 | 0.9584 |
| 0.1007 | 22.29 | 7400 | 0.1139 | 0.9580 | 0.9580 |
| 0.1049 | 22.89 | 7600 | 0.1084 | 0.9608 | 0.9608 |
| 0.0978 | 23.49 | 7800 | 0.1111 | 0.9606 | 0.9606 |
| 0.1003 | 24.1 | 8000 | 0.1097 | 0.9614 | 0.9614 |
| 0.0998 | 24.7 | 8200 | 0.1112 | 0.9606 | 0.9606 |
| 0.0973 | 25.3 | 8400 | 0.1105 | 0.9614 | 0.9614 |
| 0.0993 | 25.9 | 8600 | 0.1115 | 0.9599 | 0.9599 |
| 0.1001 | 26.51 | 8800 | 0.1096 | 0.9606 | 0.9606 |
| 0.0978 | 27.11 | 9000 | 0.1095 | 0.9614 | 0.9614 |
| 0.0999 | 27.71 | 9200 | 0.1091 | 0.9619 | 0.9619 |
| 0.0965 | 28.31 | 9400 | 0.1094 | 0.9621 | 0.9621 |
| 0.101 | 28.92 | 9600 | 0.1095 | 0.9614 | 0.9614 |
| 0.0945 | 29.52 | 9800 | 0.1099 | 0.9616 | 0.9616 |
| 0.1003 | 30.12 | 10000 | 0.1096 | 0.9617 | 0.9617 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_32768_512_30M", "model-index": [{"name": "GUE_prom_prom_300_notata-seqsight_32768_512_30M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_prom_prom_300_notata-seqsight_32768_512_30M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_32768_512_30M",
"region:us"
] | null | 2024-04-30T03:47:04+00:00 |
null | peft |
<!-- 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. -->
# GUE_prom_prom_300_notata-seqsight_32768_512_30M-L32_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_32768_512_30M](https://huggingface.co/mahdibaghbanzadeh/seqsight_32768_512_30M) on the [mahdibaghbanzadeh/GUE_prom_prom_300_notata](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_prom_prom_300_notata) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1260
- F1 Score: 0.9555
- Accuracy: 0.9555
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.2619 | 0.6 | 200 | 0.1383 | 0.9463 | 0.9463 |
| 0.1428 | 1.2 | 400 | 0.1264 | 0.9550 | 0.9550 |
| 0.1394 | 1.81 | 600 | 0.1199 | 0.9561 | 0.9561 |
| 0.1262 | 2.41 | 800 | 0.1168 | 0.9568 | 0.9568 |
| 0.1287 | 3.01 | 1000 | 0.1176 | 0.9561 | 0.9561 |
| 0.1213 | 3.61 | 1200 | 0.1226 | 0.9536 | 0.9536 |
| 0.1232 | 4.22 | 1400 | 0.1126 | 0.9583 | 0.9584 |
| 0.1178 | 4.82 | 1600 | 0.1121 | 0.9585 | 0.9585 |
| 0.1155 | 5.42 | 1800 | 0.1149 | 0.9576 | 0.9576 |
| 0.1099 | 6.02 | 2000 | 0.1149 | 0.9563 | 0.9563 |
| 0.1137 | 6.63 | 2200 | 0.1095 | 0.9595 | 0.9595 |
| 0.1078 | 7.23 | 2400 | 0.1109 | 0.9606 | 0.9606 |
| 0.1043 | 7.83 | 2600 | 0.1122 | 0.9595 | 0.9595 |
| 0.1062 | 8.43 | 2800 | 0.1079 | 0.9599 | 0.9599 |
| 0.1083 | 9.04 | 3000 | 0.1101 | 0.9608 | 0.9608 |
| 0.1032 | 9.64 | 3200 | 0.1136 | 0.9585 | 0.9585 |
| 0.0994 | 10.24 | 3400 | 0.1114 | 0.9599 | 0.9599 |
| 0.102 | 10.84 | 3600 | 0.1097 | 0.9593 | 0.9593 |
| 0.0956 | 11.45 | 3800 | 0.1078 | 0.9593 | 0.9593 |
| 0.0971 | 12.05 | 4000 | 0.1105 | 0.9632 | 0.9633 |
| 0.0976 | 12.65 | 4200 | 0.1063 | 0.9625 | 0.9625 |
| 0.0976 | 13.25 | 4400 | 0.1096 | 0.9610 | 0.9610 |
| 0.0938 | 13.86 | 4600 | 0.1071 | 0.9616 | 0.9616 |
| 0.0934 | 14.46 | 4800 | 0.1098 | 0.9631 | 0.9631 |
| 0.0917 | 15.06 | 5000 | 0.1059 | 0.9621 | 0.9621 |
| 0.0904 | 15.66 | 5200 | 0.1117 | 0.9600 | 0.9601 |
| 0.0926 | 16.27 | 5400 | 0.1090 | 0.9614 | 0.9614 |
| 0.0871 | 16.87 | 5600 | 0.1079 | 0.9623 | 0.9623 |
| 0.0845 | 17.47 | 5800 | 0.1060 | 0.9616 | 0.9616 |
| 0.0902 | 18.07 | 6000 | 0.1082 | 0.9623 | 0.9623 |
| 0.0847 | 18.67 | 6200 | 0.1080 | 0.9636 | 0.9636 |
| 0.0849 | 19.28 | 6400 | 0.1171 | 0.9606 | 0.9606 |
| 0.0848 | 19.88 | 6600 | 0.1135 | 0.9615 | 0.9616 |
| 0.0864 | 20.48 | 6800 | 0.1098 | 0.9636 | 0.9636 |
| 0.081 | 21.08 | 7000 | 0.1101 | 0.9633 | 0.9633 |
| 0.0825 | 21.69 | 7200 | 0.1139 | 0.9610 | 0.9610 |
| 0.0805 | 22.29 | 7400 | 0.1150 | 0.9621 | 0.9621 |
| 0.0848 | 22.89 | 7600 | 0.1077 | 0.9633 | 0.9633 |
| 0.078 | 23.49 | 7800 | 0.1143 | 0.9619 | 0.9619 |
| 0.0795 | 24.1 | 8000 | 0.1107 | 0.9625 | 0.9625 |
| 0.0794 | 24.7 | 8200 | 0.1144 | 0.9621 | 0.9621 |
| 0.0759 | 25.3 | 8400 | 0.1118 | 0.9606 | 0.9606 |
| 0.0779 | 25.9 | 8600 | 0.1133 | 0.9614 | 0.9614 |
| 0.0783 | 26.51 | 8800 | 0.1114 | 0.9610 | 0.9610 |
| 0.0754 | 27.11 | 9000 | 0.1116 | 0.9617 | 0.9617 |
| 0.0772 | 27.71 | 9200 | 0.1119 | 0.9621 | 0.9621 |
| 0.0743 | 28.31 | 9400 | 0.1120 | 0.9623 | 0.9623 |
| 0.0784 | 28.92 | 9600 | 0.1122 | 0.9625 | 0.9625 |
| 0.0729 | 29.52 | 9800 | 0.1127 | 0.9621 | 0.9621 |
| 0.0768 | 30.12 | 10000 | 0.1122 | 0.9621 | 0.9621 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_32768_512_30M", "model-index": [{"name": "GUE_prom_prom_300_notata-seqsight_32768_512_30M-L32_f", "results": []}]} | mahdibaghbanzadeh/GUE_prom_prom_300_notata-seqsight_32768_512_30M-L32_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_32768_512_30M",
"region:us"
] | null | 2024-04-30T03:49:44+00:00 |
null | null | {} | feliphe-galiza/whisper-small-hi | null | [
"region:us"
] | null | 2024-04-30T03:50:34+00:00 |
|
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | cilantro9246/be3df63 | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:51:06+00:00 |
null | peft |
<!-- 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. -->
# nash_dpo_iter_1
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the updated and the original datasets.
It achieves the following results on the evaluation set:
- Loss: 0.6285
- Rewards/chosen: -0.1131
- Rewards/rejected: -0.2857
- Rewards/accuracies: 0.7000
- Rewards/margins: 0.1725
- Logps/rejected: -286.0817
- Logps/chosen: -295.3544
- Logits/rejected: -2.5416
- Logits/chosen: -2.6261
## 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: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 16
- total_train_batch_size: 128
- total_eval_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
| 0.6297 | 0.64 | 100 | 0.6285 | -0.1131 | -0.2857 | 0.7000 | 0.1725 | -286.0817 | -295.3544 | -2.5416 | -2.6261 |
### Framework versions
- PEFT 0.7.1
- Transformers 4.36.2
- Pytorch 2.1.2+cu121
- Datasets 2.14.6
- Tokenizers 0.15.2 | {"license": "apache-2.0", "library_name": "peft", "tags": ["alignment-handbook", "generated_from_trainer", "trl", "dpo"], "datasets": ["updated", "original"], "base_model": "alignment-handbook/zephyr-7b-sft-full", "model-index": [{"name": "nash_dpo_iter_1", "results": []}]} | YYYYYYibo/nash_dpo_iter_1 | null | [
"peft",
"tensorboard",
"safetensors",
"mistral",
"alignment-handbook",
"generated_from_trainer",
"trl",
"dpo",
"dataset:updated",
"dataset:original",
"base_model:alignment-handbook/zephyr-7b-sft-full",
"license:apache-2.0",
"region:us"
] | null | 2024-04-30T03:51:13+00:00 |
text-classification | transformers | # SWOT Analysis Model based on DistilBERT
This repository hosts a fine-tuned version of `distilbert-base-uncased`, specifically trained to classify SWOT elements (Strength, Weakness, Opportunity, Threat) in Amazon product reviews of smartphones. This model serves as a "Synthetic Expert", with annotations derived from a combination of GPT-4 generated labels and human labeling.
## Model Training and Data
- **Base Model**: `distilbert-base-uncased`
- **Dataset**: 9,545 Amazon product reviews.
- **Annotations**:
- GPT-4 generated labels for 9,045 reviews.
- Human-labeled data for 500 reviews as a baseline.
- **Task**: Multi-label classification of SWOT elements.
## How to Use
This model can be directly loaded via the Hugging Face Transformers library:
```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer
# Import model and tokenizer
model = AutoModelForSequenceClassification.from_pretrained('jcaponigro/SWOT_Classifier')
tokenizer = AutoTokenizer.from_pretrained('jcaponigro/SWOT_Classifier')
# Example of model usage
text = "Your text for SWOT analysis."
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
``` | {"license": "mit"} | jcaponigro/SWOT_Classifier | null | [
"transformers",
"safetensors",
"distilbert",
"text-classification",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:51:15+00:00 |
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | nem012/gemma2b-2e-4 | null | [
"transformers",
"safetensors",
"gemma",
"text-generation",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:51:17+00:00 |
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": ["unsloth"]} | trex5790/model_l1 | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"unsloth",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"4-bit",
"region:us"
] | null | 2024-04-30T03:51:19+00:00 |
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | nem012/gemma2b-1e-3 | null | [
"transformers",
"safetensors",
"gemma",
"text-generation",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:51:37+00:00 |
text-generation | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | shallow6414/4dvvg8b | null | [
"transformers",
"safetensors",
"llama",
"text-generation",
"conversational",
"arxiv:1910.09700",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null | 2024-04-30T03:52:11+00:00 |
null | null | {} | Kankung/Kangkung | null | [
"region:us"
] | null | 2024-04-30T03:53:31+00:00 |
|
null | transformers | ## Description
The Llama3-TAIDE model was developed and built by the National Applied Research Laboratories. This version was scaled to 32k using RoPE scaling.
This repo contains GGUF format model files for [Llama3-TAIDE-LX-8B-Chat-Alpha1](https://huggingface.co/taide/Llama3-TAIDE-LX-8B-Chat-Alpha1).
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
## Provided files
| Name | Quant method | Bits | Size | Use case |
| ---- | ---- | ---- | ---- | ---- |
| [Llama3-TAIDE-LX-8B-Chat-Alpha1-32k-GGUF.Q5_K_M.gguf](https://huggingface.co/chienweichang/Llama3-TAIDE-LX-8B-Chat-Alpha1-32k-GGUF/blob/main/Llama3-TAIDE-LX-8B-Chat-Alpha1-32k-GGUF-unsloth.Q5_K_M.gguf) | Q5_K_M | 5 | 5.73 GB| large, very low quality loss |
| [Llama3-TAIDE-LX-8B-Chat-Alpha1-32k-GGUF.Q6_K.gguf](https://huggingface.co/chienweichang/Llama3-TAIDE-LX-8B-Chat-Alpha1-32k-GGUF/blob/main/Llama3-TAIDE-LX-8B-Chat-Alpha1-32k-GGUF-unsloth.Q6_K.gguf) | Q6_K | 6 | 6.6 GB| very large, extremely low quality loss |
## Original model card
---
* [English Version](https://huggingface.co/taide/Llama3-TAIDE-LX-8B-Chat-Alpha1/blob/main/README_en.md)
# 模型簡介
* [TAIDE計畫](https://taide.tw/index)致力於開發符合台灣語言和文化特性的生成式人工智慧對話引擎模型，同時建構可信任的人工智慧環境。結合產學研能量，推動可信任生成式人工智慧的發展，提升台灣在國際競爭中的地位，促進產業發展，避免對外國技術的依賴。
* Llama3 TAIDE 系列模型以 Meta 公司釋出的 [LLaMA3-8b](https://llama.meta.com/docs/model-cards-and-prompt-formats/meta-llama-3/) 為基礎，導入台灣不同領域可用的文本與訓練素材，提高模型在正體中文回應的能力與特定任務的表現。公開釋出的模型如下:
* [Llama3-TAIDE-LX-8B-Chat-Alpha1](https://huggingface.co/taide/Llama3-TAIDE-LX-8B-Chat-Alpha1): 以 LLaMA3-8b 為基礎，使用繁體中文資料預訓練 (continuous pretraining)，並透過指令微調(instruction tuning)強化辦公室常用任務和多輪問答對話能力，適合聊天對話或任務協助的使用情境。Llama3-TAIDE-LX-8B-Chat-Alpha1 另外有提供[4 bit 量化模型](https://huggingface.co/taide/Llama3-TAIDE-LX-8B-Chat-Alpha1-4bit)，量化模型主要是提供使用者的便利性，可能會影響效能與更多不可預期的問題，還請使用者理解與注意。
# 模型參數
* 參數量: 8B
* 最大內容長度 (context length): 8K
* 繁中訓練資料 token 量: 43B
* 訓練時間: 2336 H100 GPU Hours
# 特色
* 嚴格把關模型的訓練資料，提升模型生成資料的可信任性和適用性
* 針對自動摘要、寫信、寫文章、中翻英、英翻中等辦公室常用任務做加強
* 針對台灣在地文化、用語、國情等知識做加強
* 具備多輪問答對話能力
# 應用範例
| 任務 | 使用者輸入 | 模型輸出 |
| --- | -------- | ------- |
| 寫文章 | 請以以下內容為基礎，寫一篇文章：撰寫一篇作文，題目為《一張舊照片》，內容要求為：選擇一張令你印象深刻的照片，說明令你印象深刻的原因，並描述照片中的影像及背後的故事。記錄成長的過程、與他人的情景、環境變遷和美麗的景色。 | 一張舊照片，承載著時光的痕跡和回憶的溫度。我還記得那張照片是怎麼出現的，那是一張泛黃的黑白相片，裱在牆上，訴說著一段屬於我家族的故事。<br><br>那張照片中，我的祖父母站在田野間，後方是綿延的山巒，前方則是一片金黃的稻田。這幅景象令我印象深刻，因為這是我的童年回憶，也是我成長過程中最美好的風景線。<br><br>我出生在台灣鄉下，從小就生活在這樣一個純樸的環境中。每當我跟著家人在田裡工作時，就會看到這樣的美景。金黃的稻穗隨著微風搖曳，山巒則靜靜地矗立在遠方，給人一種安詳與祥和的感覺。<br><br>那張照片拍攝於民國50年代，當時台灣經濟尚未起飛，農業為主要產業。我的祖父母都是農民，他們辛勤耕種，為我們家提供溫飽。每當我看到那張照片，就會想起他們為我們所做的犧牲和努力。<br><br>除了美麗的風景外，照片中還有一段動人的故事。在我還很小的時候，家裡發生了一場大火，將我們的房子和所有財產都燒光了。幸好全家人都平安無事，但生活卻變得十分困苦。我們被迫搬到另一個村莊，在那裡重新開始生活。<br><br>那張照片也記錄了那段艱辛的歲月。雖然我們失去了許多東西，但家人之間的感情卻更加深厚。我們互相扶持，共同面對生活上的種種挑戰。<br><br>如今，我已長大成人，在都市裡工作。但每當我看到那張舊照片，我都會想起那段簡單而美好的時光。那些美麗的風景、親人的笑容，以及那段充滿挑戰卻又充滿希望的日子，都將永遠鐫刻在我的記憶深處。<br><br>這張舊照片，不僅僅是一張照片，更是一個寶貴的家族史，一個關於我們如何克服困難、珍惜現在的故事。我將把這張照片好好保存下來，讓後代也能夠認識我們的根源，並且銘記那些值得珍藏的美好回憶。 |
| 寫信 | 請以品牌經理的身份，給廣告公司的創意總監寫一封信，提出對於新產品廣告宣傳活動的創意建議。 | [您的姓名]<br>[您的職稱]<br>[您的公司名稱]<br>[郵件地址]<br>[電話號碼]<br>[發文日期]<br><br>[創意總監姓名]<br>[廣告公司名稱]<br>[廣告公司地址]<br><br>主旨:關於新產品宣傳活動創意發想，請查照。<br><br>尊敬的[創意總監姓名]，<br><br>我希望這封信能找到你忙碌中好好休息。我寫這封信，是為了分享我對於我們即將推出的新產品宣傳活動一些創意發想。我深信，這些想法能幫助我們新產品在市場上脫穎而出，吸引消費者的目光。<br><br>首先，我建議我們可以利用社群媒體平台，像是Facebook、Instagram、YouTube等，製作一系列引人入勝的短片和貼文。這些內容可以展現新產品的特色和優點，同時也能傳達出我們品牌重視環保、健康的核心價值。我們可以與網紅或意見領袖合作，讓他們試用並分享他們的使用經驗，這將有助於增加我們產品的可信度和能見度。<br><br>其次，我們可以舉辦新產品試吃/試用活動，邀請民眾親身體驗我們的新產品。我們可以在各大百貨公司、超市、以及我們的門市舉辦這些活動，讓參與者免費試吃/試用新產品，並填寫問卷回饋意見。這種互動的方式，不僅能收集到寶貴的市場反饋，還能讓民眾對我們的品牌產生正面的印象。<br><br>第三，我們可以與當地餐飲業者合作，將我們的新產品納入他們的菜單中。這將使我們的產品接觸到更廣泛的客群，同時也能證明我們的產品在烹飪和美食產業中具有價值。<br><br>第四，我們可以舉辦一場新產品發表會，邀請媒體、業界人士、以及重要客戶參加。在這場活動中，我們可以展示新產品的各項特色和優勢，並由專業廚師示範如何將新產品融入各式料理中。這將是向大眾展示我們品牌創新精神和專業能力的絕佳機會。<br><br>最後，我建議我們可以製作一系列精美的包裝和禮盒，以吸引消費者在節慶和送禮時選擇我們的產品。精緻的包裝將能提升我們產品的價值感，同時也能傳達出我們品牌對細節和品質的堅持。<br><br>我深信，這些創意發想將有助於我們新產品在市場上獲得成功。若您對這些想法有任何意見或建議，請不吝與我分享。我期待著與您合作，共同為[您的公司名稱]的成長和成功而努力。<br><br>謝謝您抽空閱讀我的信，並期待著與您進一步討論這些想法。<br><br>敬禮<br><br>[您的姓名]<br>[您的職稱]<br>[您的公司名稱] |
| 摘要 | 請將這篇文章精簡條理化:「產業創新條例第10條之2及第72條條文修正案」俗稱「台版晶片法」,針對半導體、電動車、5G等技術創新且居國際供應鏈關鍵地位公司,提供最高25%營所稅投抵優惠,企業適用要件包含當年度研發費用、研發密度達一定規模,且有效稅率達一定比率。<br>為因應經濟合作暨發展組織(OECD)國家最低稅負制調整,其中有效稅率門檻,民國112年訂為12%,113年料將提高至15%,但仍得審酌國際間最低稅負制實施情形。<br>經濟部官員表示,已和財政部協商進入最後階段,除企業研發密度訂在6%,目前已確認,企業購置先進製程的設備投資金額達100億元以上可抵減。<br>財政部官員表示,研商過程中,針對台灣產業與其在國際間類似的公司進行深入研究,在設備部分,畢竟適用產創10之2的業者是代表台灣隊打「國際盃」,投入金額不達100億元,可能也打不了。<br>至於備受關注的研發費用門檻,經濟部官員表示,歷經與財政部來回密切討論,研發費用門檻有望落在60億至70億元之間。<br>財政部官員指出,研發攸關台灣未來經濟成長動能,門檻不能「高不可攀」,起初雖設定在100億元,之所以會調降,正是盼讓企業覺得有辦法達得到門檻、進而適用租稅優惠,才有動力繼續投入研發,維持國際供應鏈關鍵地位。<br>經濟部官員表示,因廠商研發費用平均為30、40億元,其中,IC設計業者介於30億至60億元範圍,若將門檻訂在100億元,符合條件的業者較少、刺激誘因不足;此外,若符合申請門檻的業者增加,將可提高企業在台投資金額,財政部稅收也能因此獲得挹注。<br>IC設計業者近日頻頻針對產創10之2發聲,希望降低適用門檻,加上各國力拚供應鏈自主化、加碼補助半導體產業,經濟部官員表示,經濟部和財政部就產創10之2達成共識,爭取讓更多業者受惠,盼增強企業投資力道及鞏固台灣技術地位。<br>財政部官員表示,租稅獎勵的制定必須「有為有守」,並以達到獎勵設置目的為最高原則,現階段在打「國內盃」的企業仍可適用產創第10條、10之1的租稅優惠,共同壯大台灣經濟發展。<br>經濟部和財政部正就研發費用門檻做最後確認,待今明兩天預告子法之後,約有30天時間,可與業界進一步討論及調整,盼產創10之2能在6月上路。 | 文章重點摘要如下:<br><br>- 「台版晶片法」或《產業創新條例第10條之2及第72條條文修正案》，旨在鼓勵關鍵技術創新企業，特別是那些在半導體、電動車、5G技術等領域且在國際供應鏈中扮演關鍵角色者。<br><br>- 企業若符合特定條件，包括當年度研發費用及研發密度達標，可享最高25%的營所稅抵減優惠。研發密度目標設為6%，而企業為取得抵減，必須證明其有效稅率達到特定比率，起初設定為112年的12%，113年提高至15%，但可能會依國際稅制的實施情形進行調整。<br><br>- 政府在制定政策時，考量到OECD最低稅負制的調整，因此會視情況調整有效稅率門檻。<br><br>- 政府部門已就法案內容與企業利益相關者進行協商，確定一些細節，例如企業投資先進製程設備的門檻設為100億元新台幣以供抵減。<br><br>- 研發費用門檻經過討論，可能從最初提議的100億元下調至60億至70億元之間，以激勵更多企業符合資格並投入研發，維持台灣在國際供應鏈中的關鍵地位。<br><br>- 經濟部官員表示，調降門檻的目的是讓更多企業覺得能夠達成標準，從而獲得應用租稅優惠的動力，並可能藉此提高在台投資金額，進而增加政府稅收。<br><br>- 由於IC設計業者的研發費用普遍較低，若門檻設得過高，符合條件的企業將很少，因此有必要調整。<br><br>- 在各國推動供應鏈自主化並增加對半導體產業的補助之際，經濟部正努力確保更多企業能從《產業創新條例》第10條之2中受益，以增強投資和鞏固台灣的技術地位。<br><br>- 政府強調租稅優惠的制定應符合獎勵目標，在此同時，也將繼續提供其他租稅優惠給在國內經營的企業，以促進台灣經濟成長。<br><br>- 最終的研發費用門檻將由經濟部和財政部共同確定，在預告相關子法後，將有約30天的時間供各界討論並提出調整建議，目標是讓這項新政策在6月上路。 |
| 翻譯 | 以下提供英文內容，請幫我翻譯成中文。Dongshan coffee is famous for its unique position, and the constant refinement of production methods. The flavor is admired by many caffeine afficionados. | 東山咖啡以獨特的位置，加上不斷精進的製作手法而聞名，風味更是讓許多咖啡癡趨之若鶩。 |
# 使用說明
* 範例程式
* 範例程式與文件後續會釋出到github
* prompt 樣版
* 一般問答用法
```python
chat = [
{"role": "user", "content": "{question}"},
]
prompt = tokenizer.apply_chat_template(chat)
```
* 將 {question} 替換成使用者的輸入
* 加入 system prompt 的用法
```python
chat = [
{"role": "system", "content": "{sys}"},
{"role": "user", "content": "{question}"},
]
prompt = tokenizer.apply_chat_template(chat)
```
* 將 {sys} 替換成指令，例如：你是一個來自台灣的AI助理，你的名字是 TAIDE，樂於以台灣人的立場幫助使用者，會用繁體中文回答問題。
* 將 {question} 替換成使用者的問題
* 多輪問答用法
```python
chat = [
{"role": "system", "content": "{sys}"},
{"role": "user", "content": "{question1}"},
{"role": "assistant", "content": "{model_anwer_1}"},
{"role": "user", "content": "{question2}"},
]
prompt = tokenizer.apply_chat_template(chat)
```
* 將 {sys} 替換成指令，例如：你是一個來自台灣的AI助理，你的名字是 TAIDE，樂於以台灣人的立場幫助使用者，會用繁體中文回答問題。
* 將 {question1} 替換成使用者的問題1
* 將 {model_anwer_1} 替換成模型的回答1
* 將 {question2} 替換成使用者的問題2
* 更多細節請參考[Llama3 文件](https://llama.meta.com/docs/model-cards-and-prompt-formats/meta-llama-3/)
# 訓練方法
* 軟硬體規格
* 國網中心 H100
* 訓練框架: PyTorch
* 資料前處理
* 字元標準化
* 去除重覆
* 去除雜訊
* 網頁資料的html tag、javascript
* 非標準字元或亂碼
* 字數過短的文章
* 去除文章中的特定格式，如為排版增加的換行
* 去除個資，如email、電話
* 去除不當文字，如賭博、色情等
* 持續預訓練 (continuous pretraining, CP)
* 補充大量來源可信賴的繁體中文知識
* 超參數 (hyper parameters)
* optimizer: AdamW
* learning rate: 1e-4
* batch size: 1M tokens
* epoch: 1
* 微調 (fine tune, FT)
* 讓模型可針對繁體中文提問回答問題
* 超參數 (hyper parameters)
* optimizer: AdamW
* learning rate: 5e-5
* batch size: 256K tokens
* epoch: 3
# 訓練資料
* 持續預訓練資料(資料量約為140G)
| 資料集 | 資料描述 |
| --- | -------- |
| 訴訟資料 | 《[司法院裁判書](https://judgment.judicial.gov.tw/FJUD/default.aspx)》自2013年1月至2023年12月各級法院民事、刑事、行政訴訟資料。 |
| 中央社 | 《[中央社中文新聞](https://www.cna.com.tw/)》資料集含中央社自1993年6月至2023年06月，共30年份之每日新聞文章。內容涵蓋國內外政治、社會、財經、文教、生活等領域。 |
| ETtoday 新聞雲 | 《[ETtoday新聞雲](https://www.ettoday.net/)》資料，包含自2011年10月至 2023年12月的資料。 |
| 立法院公報 | 《[立法院公報](https://ppg.ly.gov.tw/ppg/)》包含自第8屆第1會期至第10屆第7會期之公報資料。 |
| 出版商網站書籍介紹 | 包含[三采](https://www.suncolor.com.tw/)、[Gotop](https://www.gotop.com.tw/)出版商網站上的書籍簡介。 |
| GRB 研究計畫摘要 | [GRB](https://www.grb.gov.tw/)為收錄由政府經費補助之研究計畫及其成果報告的資訊系統，此資料集主要收錄 1993年至 2023年之研究計畫摘要以及研究報告摘要，含中文及其英文對照。 |
| 學術會議論文摘要 | 收錄《[學術會議論文摘要資料庫](https://sticnet.stpi.narl.org.tw/sticloc/ttscalle?meet:)》中自1988至2009年由台灣所舉辦之學術會議論文。 |
| 光華雜誌 | 《[台灣光華雜誌](https://www.taiwan-panorama.com/)》含自1993年7月至2023年6月的文章，共30年份。內容著重於我國文化、觀光與民情等。 |
| 樂詞網 | 《[樂詞網](https://terms.naer.edu.tw/)》涵蓋文理領域約187萬則學術名詞及其譯名對照。 |
| 各部會資料 | 包含行政院「[國情簡介](https://www.ey.gov.tw/state/)」、文化部「[國家文化記憶庫](https://memory.culture.tw/)」、國發會「[檔案支援教學網](https://art.archives.gov.tw/index.aspx)」、交通部「[交通安全入口網](https://168.motc.gov.tw/)」等部會網站資料之部分資料。 |
| 今周刊 | 《[今周刊](https://www.businesstoday.com.tw/)》為一以財經為主的週刊雜誌，此資料集涵蓋2008年1月至2023年7月的文章。 |
| 教育部國語辭典、成語辭典 | 包含以下三項資料:<br>[教育部《成語典》](https://dict.idioms.moe.edu.tw/search.jsp?webMd=1&la=0)，含5,338條成語，內容包含每條成語的釋義、典故原文及其白話說明、用法說明、例句等。<br>[教育部《重編國語辭典修訂本》](https://dict.revised.moe.edu.tw/?la=0&powerMode=0)，收錄中文單字及各類辭彙，包含讀音、部首、釋義等資訊，共約165,539筆資料。<br>[教育部《國語辭典簡編本》](https://dict.concised.moe.edu.tw/?la=0&powerMode=0)，為《重編國語辭典修訂本》的簡編版本，共45,247筆資料。 |
| 科技大觀園資料 | 含《[科技大觀園網站](https://scitechvista.nat.gov.tw/)》上的科學新知以及科普文章。 |
| iKnow 科技產業資訊室 | 《[科技產業資訊室](https://iknow.stpi.narl.org.tw/)（iKnow）》提供台灣及全球的科技市場趨勢、策略分析、專利知識，及技術交易資訊，專注於科技產業的創新與發展，包含自 2008 年至 2023 年。 |
| 科學發展月刊 | 《[科學發展月刊](https://ejournal.stpi.narl.org.tw/sd)》為國科會為推廣科學教育而出版的科普刊物，含自2004年10月至2020年12月之科普文章；2021年起，以《[科技魅癮](https://www.charmingscitech.nat.gov.tw/)》季刊重新出發，提供國際關注科技議題的新知文章。 |
| 法規資料庫 | 《[法規資料庫](https://law.moj.gov.tw/)》含截自 112 年 10 月各政府部門最新發布之中央法規、行政規則、法規命令草案及地方自治法規等。 |
| 各地政府旅遊網 | 涵蓋台灣部分縣市地方政府觀光旅遊網站上之部分資料。 |
| 國教院課程綱要(十二年國教) | 含十二年國教課程綱要之總綱以及各級學校不同科目之課程綱要。 |
| 中央社譯名檔資料庫 | 《中央社譯名檔資料庫》蒐集中央社新聞業務上翻譯過的中外姓氏、人名、組織、地名等譯名對照。 |
| 童話書 | 共 20 本童話書，含湯姆歷險記、小飛俠、愛麗絲夢遊仙境、長腿叔叔等。 |
| RedPajama-Data-V2 | 從國外開放多國語言語料庫 [RedPajama-Data-v2](https://github.com/togethercomputer/RedPajama-Data) 取出英文資料 |
| MathPile-commercial | 國外開放數學語料庫 [MathPile-commercial](https://huggingface.co/datasets/GAIR/MathPile_Commercial) |
| 中文維基百科 | 《[中文維基百科](https://zh.wikipedia.org/zh-tw/%E4%B8%AD%E6%96%87%E7%BB%B4%E5%9F%BA%E7%99%BE%E7%A7%91)》截至2023年1月所有條目的內容。 |
| github-code-clean | 為 github 開源程式碼資料集，去除unlicense的程式碼和文件。 |
* 微調資料
* TAIDE團隊訓練llama2系列模型來產生微調資料資料，產生的任務包含世界知識、創意寫作、普通常識、翻譯、摘要、程式、台灣價值等單輪或多輪對話問答共 128K 筆。微調資料後續會對外釋出。
# 模型評測
* taide-bench
* 評測資料
* 寫文章、寫信、摘要、英翻中、中翻英，共500題
* 資料連結: [taide-bench](https://huggingface.co/datasets/taide/taide-bench)
* 評測方法
* gpt4評分
* 評分程式: [taide-bench-eval](https://github.com/taide-taiwan/taide-bench-eval)
* 評測分數
| 模型 | 中翻英 | 英翻中 | 摘要 | 寫文章 | 寫信 | 平均 |
| --- | ----- | ----- | ---- | ---- | ---- | --- |
| Llama3-TAIDE-LX-8B-Chat-Alpha1 | 7.770 | 8.280 | 8.495 | 9.605 | 8.950 | 8.620 |
| GPT3.5 | 8.880 | 8.810 | 7.450 | 9.490 | 8.750 | 8.676 |
| TAIDE-LX-7B-Chat | 7.165 | 7.685 | 7.720 | 9.635 | 9.110 | 8.263 |
| LLAMA2 7B | 6.075 | 4.475 | 5.905 | 2.625 | 3.040 | 4.424 |
| LLAMA2 13B | 6.480 | 6.135 | 6.110 | 2.565 | 3.000 | 4.858 |
| LLAMA2 70B | 6.975 | 6.375 | 6.795 | 2.625 | 2.990 | 5.152 |
# 授權條款
* [Llama3-TAIDE 模型社群授權同意書](https://drive.google.com/file/d/12-Q0WWSjG0DW6CqJQm_jr5wUGRLeb-8p/view)
# 免責聲明
* LLM 模型由於設計架構的限制，以及資料難免有偏誤，語言模型的任何回應不代表 TAIDE 立場，使用前需要額外加入安全防護機制，且回應內容也可能包含不正確的資訊，使用者請勿盡信。
# 開發團隊
* [https://taide.tw/index/teamList](https://taide.tw/index/teamList)
# 相關連結
* [TAIDE官網](https://taide.tw/index)
* [TAIDE Huggingface](https://huggingface.co/taide)
* [TAIDE Github](https://github.com/taide-taiwan)
* [Kuwa AI](https://kuwaai.org/)
# Citation
* [TAIDE官網](https://taide.tw/index)
| {"language": ["en"], "license": "apache-2.0", "tags": ["text-generation-inference", "transformers", "unsloth", "llama", "gguf"], "base_model": "taide/Llama3-TAIDE-LX-8B-Chat-Alpha1"} | chienweichang/Llama3-TAIDE-LX-8B-Chat-Alpha1-32k-GGUF | null | [
"transformers",
"gguf",
"llama",
"text-generation-inference",
"unsloth",
"en",
"base_model:taide/Llama3-TAIDE-LX-8B-Chat-Alpha1",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:55:03+00:00 |
null | null | {} | XAnya/Kai | null | [
"region:us"
] | null | 2024-04-30T03:55:55+00:00 |
|
null | transformers |
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | {"library_name": "transformers", "tags": []} | HenryCai1129/adapter-llama-adapterhappy2sad-1k-search-3iter-50-0.0003 | null | [
"transformers",
"safetensors",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] | null | 2024-04-30T03:56:14+00:00 |
null | peft |
<!-- 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. -->
# GUE_prom_prom_core_all-seqsight_32768_512_30M-L1_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_32768_512_30M](https://huggingface.co/mahdibaghbanzadeh/seqsight_32768_512_30M) on the [mahdibaghbanzadeh/GUE_prom_prom_core_all](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_prom_prom_core_all) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4358
- F1 Score: 0.7969
- Accuracy: 0.7970
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5696 | 0.54 | 200 | 0.4894 | 0.7631 | 0.7632 |
| 0.5023 | 1.08 | 400 | 0.4744 | 0.7717 | 0.7718 |
| 0.4868 | 1.62 | 600 | 0.4699 | 0.7777 | 0.7780 |
| 0.4713 | 2.16 | 800 | 0.4664 | 0.7802 | 0.7804 |
| 0.4682 | 2.7 | 1000 | 0.4599 | 0.7859 | 0.7860 |
| 0.4631 | 3.24 | 1200 | 0.4562 | 0.7878 | 0.7878 |
| 0.4576 | 3.78 | 1400 | 0.4545 | 0.7913 | 0.7914 |
| 0.4497 | 4.32 | 1600 | 0.4558 | 0.7881 | 0.7882 |
| 0.4517 | 4.86 | 1800 | 0.4529 | 0.7916 | 0.7917 |
| 0.4492 | 5.41 | 2000 | 0.4531 | 0.7936 | 0.7936 |
| 0.444 | 5.95 | 2200 | 0.4527 | 0.7926 | 0.7926 |
| 0.4482 | 6.49 | 2400 | 0.4482 | 0.7943 | 0.7943 |
| 0.4401 | 7.03 | 2600 | 0.4495 | 0.7924 | 0.7926 |
| 0.4409 | 7.57 | 2800 | 0.4502 | 0.7958 | 0.7958 |
| 0.4334 | 8.11 | 3000 | 0.4506 | 0.7935 | 0.7936 |
| 0.4351 | 8.65 | 3200 | 0.4459 | 0.7966 | 0.7966 |
| 0.4395 | 9.19 | 3400 | 0.4469 | 0.7973 | 0.7973 |
| 0.434 | 9.73 | 3600 | 0.4454 | 0.7935 | 0.7936 |
| 0.4405 | 10.27 | 3800 | 0.4447 | 0.7938 | 0.7937 |
| 0.4335 | 10.81 | 4000 | 0.4475 | 0.7969 | 0.7970 |
| 0.4345 | 11.35 | 4200 | 0.4445 | 0.7968 | 0.7968 |
| 0.4332 | 11.89 | 4400 | 0.4444 | 0.7937 | 0.7937 |
| 0.4338 | 12.43 | 4600 | 0.4450 | 0.7974 | 0.7975 |
| 0.4343 | 12.97 | 4800 | 0.4426 | 0.7971 | 0.7971 |
| 0.4338 | 13.51 | 5000 | 0.4442 | 0.7918 | 0.7921 |
| 0.4302 | 14.05 | 5200 | 0.4422 | 0.7993 | 0.7993 |
| 0.4307 | 14.59 | 5400 | 0.4436 | 0.7999 | 0.8 |
| 0.4317 | 15.14 | 5600 | 0.4446 | 0.7954 | 0.7956 |
| 0.4318 | 15.68 | 5800 | 0.4418 | 0.7981 | 0.7981 |
| 0.4276 | 16.22 | 6000 | 0.4434 | 0.7951 | 0.7951 |
| 0.434 | 16.76 | 6200 | 0.4378 | 0.7978 | 0.7978 |
| 0.4325 | 17.3 | 6400 | 0.4395 | 0.7993 | 0.7993 |
| 0.4295 | 17.84 | 6600 | 0.4412 | 0.7961 | 0.7961 |
| 0.4264 | 18.38 | 6800 | 0.4407 | 0.7945 | 0.7946 |
| 0.4259 | 18.92 | 7000 | 0.4389 | 0.7961 | 0.7961 |
| 0.4253 | 19.46 | 7200 | 0.4409 | 0.7973 | 0.7973 |
| 0.4327 | 20.0 | 7400 | 0.4398 | 0.7966 | 0.7966 |
| 0.4293 | 20.54 | 7600 | 0.4412 | 0.7968 | 0.7968 |
| 0.4253 | 21.08 | 7800 | 0.4409 | 0.7966 | 0.7966 |
| 0.4259 | 21.62 | 8000 | 0.4393 | 0.7978 | 0.7978 |
| 0.4254 | 22.16 | 8200 | 0.4396 | 0.7998 | 0.7998 |
| 0.4309 | 22.7 | 8400 | 0.4376 | 0.7961 | 0.7961 |
| 0.4247 | 23.24 | 8600 | 0.4387 | 0.7981 | 0.7981 |
| 0.4262 | 23.78 | 8800 | 0.4391 | 0.7987 | 0.7986 |
| 0.4233 | 24.32 | 9000 | 0.4399 | 0.7990 | 0.7990 |
| 0.4257 | 24.86 | 9200 | 0.4392 | 0.7981 | 0.7981 |
| 0.4327 | 25.41 | 9400 | 0.4382 | 0.7971 | 0.7971 |
| 0.4205 | 25.95 | 9600 | 0.4392 | 0.7981 | 0.7981 |
| 0.4292 | 26.49 | 9800 | 0.4388 | 0.7970 | 0.7970 |
| 0.4231 | 27.03 | 10000 | 0.4390 | 0.7973 | 0.7973 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_32768_512_30M", "model-index": [{"name": "GUE_prom_prom_core_all-seqsight_32768_512_30M-L1_f", "results": []}]} | mahdibaghbanzadeh/GUE_prom_prom_core_all-seqsight_32768_512_30M-L1_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_32768_512_30M",
"region:us"
] | null | 2024-04-30T03:56:55+00:00 |
null | peft |
<!-- 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. -->
# GUE_prom_prom_core_all-seqsight_32768_512_30M-L8_f
This model is a fine-tuned version of [mahdibaghbanzadeh/seqsight_32768_512_30M](https://huggingface.co/mahdibaghbanzadeh/seqsight_32768_512_30M) on the [mahdibaghbanzadeh/GUE_prom_prom_core_all](https://huggingface.co/datasets/mahdibaghbanzadeh/GUE_prom_prom_core_all) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4155
- F1 Score: 0.8044
- Accuracy: 0.8044
## 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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Score | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.542 | 0.54 | 200 | 0.4780 | 0.7712 | 0.7715 |
| 0.4765 | 1.08 | 400 | 0.4652 | 0.7851 | 0.7853 |
| 0.4561 | 1.62 | 600 | 0.4509 | 0.7936 | 0.7936 |
| 0.4453 | 2.16 | 800 | 0.4528 | 0.7968 | 0.7968 |
| 0.444 | 2.7 | 1000 | 0.4474 | 0.7930 | 0.7931 |
| 0.441 | 3.24 | 1200 | 0.4422 | 0.7939 | 0.7939 |
| 0.4373 | 3.78 | 1400 | 0.4479 | 0.7863 | 0.7870 |
| 0.43 | 4.32 | 1600 | 0.4444 | 0.7928 | 0.7929 |
| 0.4332 | 4.86 | 1800 | 0.4427 | 0.7925 | 0.7927 |
| 0.4301 | 5.41 | 2000 | 0.4451 | 0.7901 | 0.7902 |
| 0.4273 | 5.95 | 2200 | 0.4423 | 0.7990 | 0.7990 |
| 0.4309 | 6.49 | 2400 | 0.4358 | 0.7954 | 0.7954 |
| 0.4236 | 7.03 | 2600 | 0.4401 | 0.7978 | 0.7980 |
| 0.4259 | 7.57 | 2800 | 0.4399 | 0.7975 | 0.7975 |
| 0.4179 | 8.11 | 3000 | 0.4411 | 0.7939 | 0.7941 |
| 0.4186 | 8.65 | 3200 | 0.4338 | 0.7978 | 0.7978 |
| 0.4224 | 9.19 | 3400 | 0.4376 | 0.7982 | 0.7983 |
| 0.4178 | 9.73 | 3600 | 0.4398 | 0.7911 | 0.7916 |
| 0.4237 | 10.27 | 3800 | 0.4356 | 0.7977 | 0.7978 |
| 0.4157 | 10.81 | 4000 | 0.4424 | 0.7974 | 0.7976 |
| 0.416 | 11.35 | 4200 | 0.4364 | 0.7984 | 0.7985 |
| 0.4178 | 11.89 | 4400 | 0.4376 | 0.7955 | 0.7959 |
| 0.4158 | 12.43 | 4600 | 0.4348 | 0.7989 | 0.7990 |
| 0.4162 | 12.97 | 4800 | 0.4339 | 0.7994 | 0.7995 |
| 0.4149 | 13.51 | 5000 | 0.4397 | 0.7909 | 0.7916 |
| 0.4127 | 14.05 | 5200 | 0.4313 | 0.8019 | 0.8019 |
| 0.4131 | 14.59 | 5400 | 0.4329 | 0.7996 | 0.7997 |
| 0.4125 | 15.14 | 5600 | 0.4360 | 0.8007 | 0.8008 |
| 0.4125 | 15.68 | 5800 | 0.4354 | 0.7960 | 0.7963 |
| 0.4081 | 16.22 | 6000 | 0.4357 | 0.7976 | 0.7978 |
| 0.4148 | 16.76 | 6200 | 0.4289 | 0.7996 | 0.7997 |
| 0.4124 | 17.3 | 6400 | 0.4306 | 0.8013 | 0.8014 |
| 0.4102 | 17.84 | 6600 | 0.4321 | 0.8015 | 0.8015 |
| 0.4057 | 18.38 | 6800 | 0.4365 | 0.7992 | 0.7997 |
| 0.4076 | 18.92 | 7000 | 0.4296 | 0.7999 | 0.8 |
| 0.4054 | 19.46 | 7200 | 0.4317 | 0.7982 | 0.7983 |
| 0.4121 | 20.0 | 7400 | 0.4295 | 0.8003 | 0.8003 |
| 0.4066 | 20.54 | 7600 | 0.4345 | 0.7992 | 0.7993 |
| 0.4054 | 21.08 | 7800 | 0.4318 | 0.8020 | 0.8020 |
| 0.4064 | 21.62 | 8000 | 0.4285 | 0.8027 | 0.8027 |
| 0.4042 | 22.16 | 8200 | 0.4292 | 0.8012 | 0.8012 |
| 0.4093 | 22.7 | 8400 | 0.4300 | 0.7997 | 0.7998 |
| 0.4036 | 23.24 | 8600 | 0.4290 | 0.8000 | 0.8 |
| 0.4039 | 23.78 | 8800 | 0.4301 | 0.8002 | 0.8002 |
| 0.4009 | 24.32 | 9000 | 0.4300 | 0.8019 | 0.8019 |
| 0.4032 | 24.86 | 9200 | 0.4291 | 0.8005 | 0.8005 |
| 0.4106 | 25.41 | 9400 | 0.4286 | 0.8007 | 0.8007 |
| 0.4008 | 25.95 | 9600 | 0.4294 | 0.8013 | 0.8014 |
| 0.4062 | 26.49 | 9800 | 0.4295 | 0.8003 | 0.8003 |
| 0.399 | 27.03 | 10000 | 0.4295 | 0.8007 | 0.8007 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2 | {"library_name": "peft", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "mahdibaghbanzadeh/seqsight_32768_512_30M", "model-index": [{"name": "GUE_prom_prom_core_all-seqsight_32768_512_30M-L8_f", "results": []}]} | mahdibaghbanzadeh/GUE_prom_prom_core_all-seqsight_32768_512_30M-L8_f | null | [
"peft",
"safetensors",
"generated_from_trainer",
"base_model:mahdibaghbanzadeh/seqsight_32768_512_30M",
"region:us"
] | null | 2024-04-30T03:56:56+00:00 |
null | null | {} | ZyanLeyy/gpt2-psych_chatbot-psych_chatbot | null | [
"region:us"
] | null | 2024-04-30T03:57:03+00:00 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.