|
--- |
|
license: apache-2.0 |
|
library_name: peft |
|
tags: |
|
- generated_from_trainer |
|
base_model: TheBloke/Mistral-7B-Instruct-v0.2-GPTQ |
|
model-index: |
|
- name: Mistral-7B-Instruct-v0.2-GPTQ_retrained |
|
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. --> |
|
|
|
# Mistral-7B-Instruct-v0.2-GPTQ_retrained |
|
|
|
This model is a fine-tuned version of [TheBloke/Mistral-7B-Instruct-v0.2-GPTQ](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GPTQ) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.0791 |
|
|
|
## 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: 4 |
|
- eval_batch_size: 4 |
|
- 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_steps: 2 |
|
- num_epochs: 10 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 1.6921 | 0.92 | 3 | 1.5225 | |
|
| 1.629 | 1.85 | 6 | 1.4415 | |
|
| 1.5316 | 2.77 | 9 | 1.3478 | |
|
| 1.0839 | 4.0 | 13 | 1.2525 | |
|
| 1.3848 | 4.92 | 16 | 1.2006 | |
|
| 1.3332 | 5.85 | 19 | 1.1594 | |
|
| 1.2817 | 6.77 | 22 | 1.1265 | |
|
| 0.9486 | 8.0 | 26 | 1.0928 | |
|
| 1.2404 | 8.92 | 29 | 1.0804 | |
|
| 0.8768 | 9.23 | 30 | 1.0791 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.10.0 |
|
- Transformers 4.38.2 |
|
- Pytorch 2.1.0+cu121 |
|
- Datasets 2.18.0 |
|
- Tokenizers 0.15.2 |