| --- |
| license: other |
| base_model: yahma/llama-7b-hf |
| tags: |
| - generated_from_trainer |
| model-index: |
| - name: V0305O2 |
| 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. --> |
|
|
| # V0305O2 |
|
|
| This model is a fine-tuned version of [yahma/llama-7b-hf](https://huggingface.co/yahma/llama-7b-hf) on an unknown dataset. |
| It achieves the following results on the evaluation set: |
| - Loss: 0.1482 |
|
|
| ## 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: 0.0003 |
| - train_batch_size: 4 |
| - eval_batch_size: 8 |
| - seed: 42 |
| - gradient_accumulation_steps: 32 |
| - total_train_batch_size: 128 |
| - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
| - lr_scheduler_type: cosine_with_restarts |
| - lr_scheduler_warmup_steps: 20 |
| - num_epochs: 3 |
| - mixed_precision_training: Native AMP |
|
|
| ### Training results |
|
|
| | Training Loss | Epoch | Step | Validation Loss | |
| |:-------------:|:-----:|:----:|:---------------:| |
| | 3.7252 | 0.09 | 10 | 0.2911 | |
| | 0.1739 | 0.17 | 20 | 0.1561 | |
| | 0.1571 | 0.26 | 30 | 0.1529 | |
| | 0.1516 | 0.34 | 40 | 0.1509 | |
| | 0.1514 | 0.43 | 50 | 0.1507 | |
| | 0.1564 | 0.51 | 60 | 0.1511 | |
| | 0.1519 | 0.6 | 70 | 0.1509 | |
| | 0.1534 | 0.68 | 80 | 0.1493 | |
| | 0.1492 | 0.77 | 90 | 0.1490 | |
| | 0.1527 | 0.85 | 100 | 0.1498 | |
| | 0.1543 | 0.94 | 110 | 0.1491 | |
| | 0.1504 | 1.02 | 120 | 0.1506 | |
| | 0.1536 | 1.11 | 130 | 0.1498 | |
| | 0.1493 | 1.19 | 140 | 0.1504 | |
| | 0.1526 | 1.28 | 150 | 0.1492 | |
| | 0.1529 | 1.37 | 160 | 0.1500 | |
| | 0.1519 | 1.45 | 170 | 0.1490 | |
| | 0.1491 | 1.54 | 180 | 0.1495 | |
| | 0.1539 | 1.62 | 190 | 0.1491 | |
| | 0.153 | 1.71 | 200 | 0.1485 | |
| | 0.1514 | 1.79 | 210 | 0.1492 | |
| | 0.154 | 1.88 | 220 | 0.1501 | |
| | 0.1562 | 1.96 | 230 | 0.1490 | |
| | 0.1511 | 2.05 | 240 | 0.1491 | |
| | 0.154 | 2.13 | 250 | 0.1484 | |
| | 0.1505 | 2.22 | 260 | 0.1489 | |
| | 0.1492 | 2.3 | 270 | 0.1488 | |
| | 0.1546 | 2.39 | 280 | 0.1478 | |
| | 0.1528 | 2.47 | 290 | 0.1476 | |
| | 0.1516 | 2.56 | 300 | 0.1479 | |
| | 0.1508 | 2.65 | 310 | 0.1482 | |
| | 0.1523 | 2.73 | 320 | 0.1482 | |
| | 0.1501 | 2.82 | 330 | 0.1483 | |
| | 0.1498 | 2.9 | 340 | 0.1483 | |
| | 0.151 | 2.99 | 350 | 0.1482 | |
|
|
|
|
| ### Framework versions |
|
|
| - Transformers 4.36.0.dev0 |
| - Pytorch 2.1.2+cu121 |
| - Datasets 2.14.6 |
| - Tokenizers 0.14.1 |
|
|
