| --- |
| license: mit |
| base_model: microsoft/phi-2 |
| tags: |
| - generated_from_trainer |
| model-index: |
| - name: V0316MP2 |
| 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. --> |
|
|
| # V0316MP2 |
|
|
| This model is a fine-tuned version of [microsoft/phi-2](https://huggingface.co/microsoft/phi-2) on an unknown dataset. |
| It achieves the following results on the evaluation set: |
| - Loss: 0.0962 |
|
|
| ## 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 | |
| |:-------------:|:-----:|:----:|:---------------:| |
| | 2.5399 | 0.09 | 10 | 2.3219 | |
| | 2.1738 | 0.17 | 20 | 1.8070 | |
| | 1.6126 | 0.26 | 30 | 1.2246 | |
| | 1.1047 | 0.34 | 40 | 0.7910 | |
| | 0.6789 | 0.43 | 50 | 0.3123 | |
| | 0.3195 | 0.51 | 60 | 0.1536 | |
| | 0.2157 | 0.6 | 70 | 0.1208 | |
| | 0.1791 | 0.68 | 80 | 0.1139 | |
| | 0.16 | 0.77 | 90 | 0.1100 | |
| | 0.1628 | 0.85 | 100 | 0.1076 | |
| | 0.1556 | 0.94 | 110 | 0.1066 | |
| | 0.1509 | 1.02 | 120 | 0.1057 | |
| | 0.1575 | 1.11 | 130 | 0.1040 | |
| | 0.1502 | 1.19 | 140 | 0.1038 | |
| | 0.148 | 1.28 | 150 | 0.1024 | |
| | 0.1478 | 1.37 | 160 | 0.1019 | |
| | 0.1469 | 1.45 | 170 | 0.1015 | |
| | 0.1339 | 1.54 | 180 | 0.1008 | |
| | 0.1433 | 1.62 | 190 | 0.1002 | |
| | 0.1408 | 1.71 | 200 | 0.0993 | |
| | 0.1391 | 1.79 | 210 | 0.0987 | |
| | 0.1411 | 1.88 | 220 | 0.0980 | |
| | 0.1345 | 1.96 | 230 | 0.0975 | |
| | 0.1422 | 2.05 | 240 | 0.0968 | |
| | 0.1374 | 2.13 | 250 | 0.0970 | |
| | 0.1341 | 2.22 | 260 | 0.0970 | |
| | 0.1346 | 2.3 | 270 | 0.0968 | |
| | 0.1412 | 2.39 | 280 | 0.0966 | |
| | 0.1339 | 2.47 | 290 | 0.0959 | |
| | 0.1395 | 2.56 | 300 | 0.0961 | |
| | 0.1376 | 2.65 | 310 | 0.0961 | |
| | 0.1384 | 2.73 | 320 | 0.0960 | |
| | 0.1374 | 2.82 | 330 | 0.0958 | |
| | 0.1295 | 2.9 | 340 | 0.0959 | |
| | 0.1298 | 2.99 | 350 | 0.0962 | |
|
|
|
|
| ### Framework versions |
|
|
| - Transformers 4.36.0.dev0 |
| - Pytorch 2.1.2+cu121 |
| - Datasets 2.14.6 |
| - Tokenizers 0.14.1 |
|
|
