metadata
base_model: mistralai/Mistral-7B-Instruct-v0.3
library_name: peft
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: pgd_mistral_8bits_lr9e-05_alpha32_rk4_do0.2_wd2.0e-02
results: []
pgd_mistral_8bits_lr9e-05_alpha32_rk4_do0.2_wd2.0e-02
This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.8425
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: 9e-05
- train_batch_size: 3
- eval_batch_size: 3
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 12
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 20
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.6427 | 0.9778 | 22 | 1.6811 |
| 1.0715 | 2.0 | 45 | 0.9352 |
| 0.8713 | 2.9778 | 67 | 0.8724 |
| 0.7894 | 4.0 | 90 | 0.8532 |
| 0.8064 | 4.9778 | 112 | 0.8457 |
| 0.7582 | 6.0 | 135 | 0.8422 |
| 0.7828 | 6.9778 | 157 | 0.8422 |
| 0.7403 | 8.0 | 180 | 0.8403 |
| 0.7665 | 8.9778 | 202 | 0.8410 |
| 0.7214 | 9.7778 | 220 | 0.8425 |
Framework versions
- PEFT 0.12.0
- Transformers 4.42.4
- Pytorch 2.4.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1