dev-lm-fast
This model is a fine-tuned version of microsoft/deberta-v3-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.8148
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: 5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1.0
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.7242 | 0.02 | 15000 | 1.6151 |
| 1.4916 | 0.04 | 30000 | 1.4046 |
| 1.3764 | 0.06 | 45000 | 1.2586 |
| 1.2939 | 0.08 | 60000 | 1.2500 |
| 1.2529 | 0.1 | 75000 | 1.1612 |
| 1.1913 | 0.12 | 90000 | 1.1410 |
| 1.1712 | 0.15 | 105000 | 1.0923 |
| 1.1435 | 0.17 | 120000 | 1.0421 |
| 1.1156 | 0.19 | 135000 | 1.0778 |
| 1.1105 | 0.21 | 150000 | 1.0409 |
| 1.0968 | 0.23 | 165000 | 1.0334 |
| 1.0829 | 0.25 | 180000 | 1.0151 |
| 1.0591 | 0.27 | 195000 | 0.9855 |
| 1.0606 | 0.29 | 210000 | 0.9567 |
| 1.0556 | 0.31 | 225000 | 0.9461 |
| 1.0349 | 0.33 | 240000 | 0.9628 |
| 1.0323 | 0.35 | 255000 | 0.9434 |
| 1.0055 | 0.37 | 270000 | 0.9502 |
| 1.0044 | 0.39 | 285000 | 0.9646 |
| 0.9886 | 0.42 | 300000 | 0.9216 |
| 0.9783 | 0.44 | 315000 | 0.9166 |
| 0.9646 | 0.46 | 330000 | 0.9277 |
| 0.9767 | 0.48 | 345000 | 0.8874 |
| 0.9653 | 0.5 | 360000 | 0.9212 |
| 0.95 | 0.52 | 375000 | 0.8993 |
| 0.947 | 0.54 | 390000 | 0.8900 |
| 0.9501 | 0.56 | 405000 | 0.8936 |
| 0.9505 | 0.58 | 420000 | 0.8920 |
| 0.9534 | 0.6 | 435000 | 0.8688 |
| 0.9412 | 0.62 | 450000 | 0.8820 |
| 0.9269 | 0.64 | 465000 | 0.8507 |
| 0.9237 | 0.66 | 480000 | 0.8848 |
| 0.9153 | 0.69 | 495000 | 0.8481 |
| 0.9132 | 0.71 | 510000 | 0.8403 |
| 0.8967 | 0.73 | 525000 | 0.8634 |
| 0.901 | 0.75 | 540000 | 0.8334 |
| 0.8941 | 0.77 | 555000 | 0.8441 |
| 0.9084 | 0.79 | 570000 | 0.8088 |
| 0.9238 | 0.81 | 585000 | 0.8470 |
| 0.8642 | 0.83 | 600000 | 0.8108 |
| 0.8887 | 0.85 | 615000 | 0.8180 |
| 0.8819 | 0.87 | 630000 | 0.8174 |
| 0.8807 | 0.89 | 645000 | 0.8235 |
| 0.896 | 0.91 | 660000 | 0.8298 |
| 0.8735 | 0.93 | 675000 | 0.8041 |
| 0.8616 | 0.96 | 690000 | 0.8353 |
| 0.8727 | 0.98 | 705000 | 0.7769 |
| 0.8737 | 1.0 | 720000 | 0.7964 |
Framework versions
- Transformers 4.13.0.dev0
- Pytorch 1.10.0+cu113
- Datasets 1.15.1
- Tokenizers 0.10.3
- Downloads last month
- 264