---
base_model: meta-llama/Meta-Llama-3-8B
datasets:
- generator
library_name: peft
license: llama3
tags:
- trl
- sft
- generated_from_trainer
model-index:
- name: POC-Meta-Llama-3-8B-MEDAL-flash-attention-2-cosine-evaldata
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# POC-Meta-Llama-3-8B-MEDAL-flash-attention-2-cosine-evaldata

This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on the generator dataset.
It achieves the following results on the evaluation set:
- Loss: 2.2897

## Model description

Article: https://medium.com/@frankmorales_91352/fine-tuning-meta-llama-3-8b-with-medal-a-refined-approach-for-enhanced-medical-language-b924d226b09d


## Training and evaluation data

Fine-Tuning: https://github.com/frank-morales2020/MLxDL/blob/main/FineTuning_LLM_Meta_Llama_3_8B_for_MEDAL_EVALDATA.ipynb

Evaluation: https://github.com/frank-morales2020/MLxDL/blob/main/Meta_Llama_3_8B_for_MEDAL_EVALUATOR_evaldata.ipynb


### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 1
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.15
- num_epochs: 0.3

### Training results

| Training Loss | Epoch  | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 2.7645        | 0.0069 | 100  | 2.6720          |
| 2.5917        | 0.0138 | 200  | 2.5243          |
| 2.5054        | 0.0207 | 300  | 2.4705          |
| 2.4406        | 0.0277 | 400  | 2.4379          |
| 2.4272        | 0.0346 | 500  | 2.4136          |
| 2.4171        | 0.0415 | 600  | 2.3942          |
| 2.3908        | 0.0484 | 700  | 2.3793          |
| 2.3808        | 0.0553 | 800  | 2.3664          |
| 2.3588        | 0.0622 | 900  | 2.3571          |
| 2.3595        | 0.0692 | 1000 | 2.3494          |
| 2.3411        | 0.0761 | 1100 | 2.3421          |
| 2.3308        | 0.0830 | 1200 | 2.3369          |
| 2.3358        | 0.0899 | 1300 | 2.3320          |
| 2.3295        | 0.0968 | 1400 | 2.3270          |
| 2.337         | 0.1037 | 1500 | 2.3228          |
| 2.3182        | 0.1106 | 1600 | 2.3195          |
| 2.3334        | 0.1176 | 1700 | 2.3161          |
| 2.3278        | 0.1245 | 1800 | 2.3128          |
| 2.3151        | 0.1314 | 1900 | 2.3101          |
| 2.3245        | 0.1383 | 2000 | 2.3075          |
| 2.3073        | 0.1452 | 2100 | 2.3053          |
| 2.3094        | 0.1521 | 2200 | 2.3036          |
| 2.3101        | 0.1590 | 2300 | 2.3013          |
| 2.3102        | 0.1660 | 2400 | 2.2995          |
| 2.3042        | 0.1729 | 2500 | 2.2980          |
| 2.2942        | 0.1798 | 2600 | 2.2965          |
| 2.2876        | 0.1867 | 2700 | 2.2951          |
| 2.3077        | 0.1936 | 2800 | 2.2941          |
| 2.2851        | 0.2005 | 2900 | 2.2931          |
| 2.2766        | 0.2075 | 3000 | 2.2923          |
| 2.2873        | 0.2144 | 3100 | 2.2916          |
| 2.2971        | 0.2213 | 3200 | 2.2910          |
| 2.2942        | 0.2282 | 3300 | 2.2906          |
| 2.2872        | 0.2351 | 3400 | 2.2903          |
| 2.2996        | 0.2420 | 3500 | 2.2901          |
| 2.2855        | 0.2489 | 3600 | 2.2899          |
| 2.2969        | 0.2559 | 3700 | 2.2898          |
| 2.2871        | 0.2628 | 3800 | 2.2898          |
| 2.2905        | 0.2697 | 3900 | 2.2897          |
| 2.2915        | 0.2766 | 4000 | 2.2897          |
| 2.2921        | 0.2835 | 4100 | 2.2897          |
| 2.3087        | 0.2904 | 4200 | 2.2897          |
| 2.3017        | 0.2974 | 4300 | 2.2897          |


### Framework versions

- PEFT 0.11.1
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1