metadata
library_name: peft
tags:
- generated_from_trainer
base_model: Narkantak/Mistral-2x7b-Instruct-1x2
model-index:
- name: working
results: []
working
This model is a fine-tuned version of Narkantak/Mistral-2x7b-Instruct-1x2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.3777
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: 6
- eval_batch_size: 6
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 24
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 2
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 3.0 | 0.96 | 12 | 0.8713 |
| 0.5738 | 2.0 | 25 | 0.4515 |
| 0.3676 | 2.96 | 37 | 0.3039 |
| 0.2127 | 4.0 | 50 | 0.2680 |
| 0.1891 | 4.96 | 62 | 0.2556 |
| 0.1427 | 6.0 | 75 | 0.2721 |
| 0.132 | 6.96 | 87 | 0.2762 |
| 0.1037 | 8.0 | 100 | 0.2844 |
| 0.1006 | 8.96 | 112 | 0.2951 |
| 0.0832 | 10.0 | 125 | 0.3044 |
| 0.0829 | 10.96 | 137 | 0.3194 |
| 0.073 | 12.0 | 150 | 0.3282 |
| 0.0757 | 12.96 | 162 | 0.3415 |
| 0.0692 | 14.0 | 175 | 0.3230 |
| 0.0738 | 14.96 | 187 | 0.3427 |
| 0.0666 | 16.0 | 200 | 0.3449 |
| 0.0712 | 16.96 | 212 | 0.3450 |
| 0.0652 | 18.0 | 225 | 0.3511 |
| 0.0698 | 18.96 | 237 | 0.3570 |
| 0.0641 | 20.0 | 250 | 0.3604 |
| 0.0691 | 20.96 | 262 | 0.3662 |
| 0.0638 | 22.0 | 275 | 0.3682 |
| 0.0689 | 22.96 | 287 | 0.3644 |
| 0.0632 | 24.0 | 300 | 0.3665 |
| 0.0681 | 24.96 | 312 | 0.3709 |
| 0.0627 | 26.0 | 325 | 0.3725 |
| 0.0677 | 26.96 | 337 | 0.3752 |
| 0.0621 | 28.0 | 350 | 0.3774 |
| 0.0645 | 28.8 | 360 | 0.3777 |
Framework versions
- PEFT 0.10.0
- Transformers 4.38.2
- Pytorch 2.1.2
- Datasets 2.1.0
- Tokenizers 0.15.2