<!-- 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. -->
MLMA_lab9_task2
This model is a fine-tuned version of microsoft/biogpt on the ncbi_disease dataset. It achieves the following results on the evaluation set:
- Loss: 1.2509
- Precision: 0.0159
- Recall: 0.1487
- F1: 0.0287
- Accuracy: 0.6365
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 1.153 | 1.0 | 680 | 1.0671 | 0.0122 | 0.1258 | 0.0223 | 0.5452 |
| 1.02 | 2.0 | 1360 | 1.0418 | 0.0098 | 0.0203 | 0.0132 | 0.6791 |
| 0.9552 | 3.0 | 2040 | 1.0269 | 0.0135 | 0.1677 | 0.0250 | 0.5282 |
| 0.926 | 4.0 | 2720 | 1.0390 | 0.0143 | 0.0940 | 0.0248 | 0.6686 |
| 0.9156 | 5.0 | 3400 | 1.0200 | 0.0135 | 0.2046 | 0.0253 | 0.4679 |
| 0.8791 | 6.0 | 4080 | 1.0543 | 0.0131 | 0.2745 | 0.0250 | 0.3149 |
| 0.8672 | 7.0 | 4760 | 1.0545 | 0.0141 | 0.2732 | 0.0267 | 0.3471 |
| 0.8627 | 8.0 | 5440 | 1.0734 | 0.0145 | 0.0826 | 0.0246 | 0.7220 |
| 0.8375 | 9.0 | 6120 | 1.1068 | 0.0156 | 0.1410 | 0.0281 | 0.6451 |
| 0.8235 | 10.0 | 6800 | 1.0796 | 0.0158 | 0.1537 | 0.0286 | 0.6210 |
| 0.8157 | 11.0 | 7480 | 1.1476 | 0.0143 | 0.1690 | 0.0263 | 0.5737 |
| 0.7957 | 12.0 | 8160 | 1.1369 | 0.0143 | 0.1525 | 0.0262 | 0.6155 |
| 0.7937 | 13.0 | 8840 | 1.2014 | 0.0151 | 0.1741 | 0.0278 | 0.5808 |
| 0.7765 | 14.0 | 9520 | 1.2249 | 0.0160 | 0.1449 | 0.0289 | 0.6443 |
| 0.7661 | 15.0 | 10200 | 1.2509 | 0.0159 | 0.1487 | 0.0287 | 0.6365 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3
