metadata
license: apache-2.0
library_name: peft
tags:
- trl
- sft
- generated_from_trainer
base_model: mistralai/Mistral-7B-v0.1
model-index:
- name: flash_attention_fine-tuned_mistral
results: []
flash_attention_fine-tuned_mistral
This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.1622
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: 6
- total_train_batch_size: 24
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.01
- training_steps: 100
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| No log | 0.0244 | 10 | 1.2046 |
| No log | 0.0487 | 20 | 1.1875 |
| 1.1978 | 0.0731 | 30 | 1.1797 |
| 1.1978 | 0.0975 | 40 | 1.1763 |
| 1.1462 | 0.1219 | 50 | 1.1736 |
| 1.1462 | 0.1462 | 60 | 1.1712 |
| 1.1462 | 0.1706 | 70 | 1.1701 |
| 1.137 | 0.1950 | 80 | 1.1681 |
| 1.137 | 0.2193 | 90 | 1.1645 |
| 1.1484 | 0.2437 | 100 | 1.1622 |
Framework versions
- PEFT 0.11.1
- Transformers 4.41.1
- Pytorch 2.2.0
- Datasets 2.19.1
- Tokenizers 0.19.1