<!-- 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. -->
chinese-macbert-base-finetuned
This model is a fine-tuned version of hfl/chinese-macbert-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2790
- Accuracy: 0.9613
- F1: 0.9548
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: 2e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|---|---|---|---|---|---|
| 1.9986 | 1.0 | 3 | 1.5298 | 0.8177 | 0.7357 |
| 1.43 | 2.0 | 6 | 1.2166 | 0.8177 | 0.7357 |
| 1.2494 | 3.0 | 9 | 1.0037 | 0.8177 | 0.7357 |
| 0.9698 | 4.0 | 12 | 0.8538 | 0.8177 | 0.7357 |
| 0.8999 | 5.0 | 15 | 0.7562 | 0.8453 | 0.7811 |
| 0.8945 | 6.0 | 18 | 0.6813 | 0.8619 | 0.7980 |
| 0.7059 | 7.0 | 21 | 0.6101 | 0.8619 | 0.7980 |
| 0.6066 | 8.0 | 24 | 0.5422 | 0.8619 | 0.7983 |
| 0.5938 | 9.0 | 27 | 0.4891 | 0.8619 | 0.7996 |
| 0.5995 | 10.0 | 30 | 0.4493 | 0.8674 | 0.8124 |
| 0.493 | 11.0 | 33 | 0.4113 | 0.8840 | 0.8470 |
| 0.5175 | 12.0 | 36 | 0.3798 | 0.9116 | 0.8893 |
| 0.4615 | 13.0 | 39 | 0.3536 | 0.9392 | 0.9218 |
| 0.4339 | 14.0 | 42 | 0.3311 | 0.9337 | 0.9094 |
| 0.3926 | 15.0 | 45 | 0.3147 | 0.9448 | 0.9317 |
| 0.3507 | 16.0 | 48 | 0.3012 | 0.9503 | 0.9384 |
| 0.3634 | 17.0 | 51 | 0.2926 | 0.9558 | 0.9471 |
| 0.2825 | 18.0 | 54 | 0.2857 | 0.9613 | 0.9548 |
| 0.308 | 19.0 | 57 | 0.2808 | 0.9613 | 0.9548 |
| 0.3323 | 20.0 | 60 | 0.2790 | 0.9613 | 0.9548 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.12.0+cu102
- Datasets 1.18.4
- Tokenizers 0.12.1
