distilbert-base-uncased_fold_3_ternary_v1
This model is a fine-tuned version of distilbert-base-uncased on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.8908
- F1: 0.7879
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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 25
Training results
| Training Loss | Epoch | Step | Validation Loss | F1 |
|---|---|---|---|---|
| No log | 1.0 | 289 | 0.5873 | 0.7636 |
| 0.5479 | 2.0 | 578 | 0.5788 | 0.7697 |
| 0.5479 | 3.0 | 867 | 0.6286 | 0.7770 |
| 0.2412 | 4.0 | 1156 | 0.8845 | 0.7661 |
| 0.2412 | 5.0 | 1445 | 0.9894 | 0.7818 |
| 0.1191 | 6.0 | 1734 | 1.0856 | 0.7842 |
| 0.0543 | 7.0 | 2023 | 1.2852 | 0.7830 |
| 0.0543 | 8.0 | 2312 | 1.4295 | 0.7673 |
| 0.0223 | 9.0 | 2601 | 1.4716 | 0.7806 |
| 0.0223 | 10.0 | 2890 | 1.6007 | 0.7636 |
| 0.0122 | 11.0 | 3179 | 1.6744 | 0.7673 |
| 0.0122 | 12.0 | 3468 | 1.6954 | 0.7685 |
| 0.0129 | 13.0 | 3757 | 1.7273 | 0.7733 |
| 0.0057 | 14.0 | 4046 | 1.7114 | 0.7758 |
| 0.0057 | 15.0 | 4335 | 1.7480 | 0.7733 |
| 0.0045 | 16.0 | 4624 | 1.8322 | 0.7830 |
| 0.0045 | 17.0 | 4913 | 1.7448 | 0.7830 |
| 0.0047 | 18.0 | 5202 | 1.8126 | 0.7782 |
| 0.0047 | 19.0 | 5491 | 1.9021 | 0.7673 |
| 0.0018 | 20.0 | 5780 | 1.9011 | 0.7830 |
| 0.0026 | 21.0 | 6069 | 1.8771 | 0.7806 |
| 0.0026 | 22.0 | 6358 | 1.8634 | 0.7806 |
| 0.0012 | 23.0 | 6647 | 1.8926 | 0.7830 |
| 0.0012 | 24.0 | 6936 | 1.8922 | 0.7855 |
| 0.0005 | 25.0 | 7225 | 1.8908 | 0.7879 |
Framework versions
- Transformers 4.21.0
- Pytorch 1.12.0+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
- Downloads last month
- 12
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.