metadata
license: apache-2.0
library_name: peft
tags:
- trl
- sft
- unsloth
- generated_from_trainer
base_model: unsloth/mistral-7b-instruct-v0.2-bnb-4bit
model-index:
- name: mistral-inst_10000_200
results: []
mistral-inst_10000_200
This model is a fine-tuned version of unsloth/mistral-7b-instruct-v0.2-bnb-4bit on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.3962
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: 8
- eval_batch_size: 4
- seed: 3407
- gradient_accumulation_steps: 8
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 5
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.4844 | 0.31 | 48 | 1.1183 |
| 1.1246 | 0.61 | 96 | 1.0908 |
| 1.1002 | 0.92 | 144 | 1.0744 |
| 0.9854 | 1.23 | 192 | 1.0764 |
| 0.9533 | 1.54 | 240 | 1.0738 |
| 0.9392 | 1.84 | 288 | 1.0642 |
| 0.8426 | 2.15 | 336 | 1.1302 |
| 0.741 | 2.46 | 384 | 1.1290 |
| 0.7343 | 2.76 | 432 | 1.1219 |
| 0.7004 | 3.07 | 480 | 1.2799 |
| 0.5405 | 3.38 | 528 | 1.2481 |
| 0.538 | 3.69 | 576 | 1.2538 |
| 0.5379 | 3.99 | 624 | 1.2606 |
| 0.4171 | 4.3 | 672 | 1.3919 |
| 0.4156 | 4.61 | 720 | 1.3977 |
| 0.4168 | 4.92 | 768 | 1.3962 |
Framework versions
- PEFT 0.10.0
- Transformers 4.39.3
- Pytorch 2.2.2+cu121
- Datasets 2.16.0
- Tokenizers 0.15.2