<!-- 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. -->
distilbert-base-uncased__hate_speech_offensive__train-32-1
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: 1.0606
- Accuracy: 0.4745
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: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.0941 | 1.0 | 19 | 1.1045 | 0.2 |
| 0.9967 | 2.0 | 38 | 1.1164 | 0.35 |
| 0.8164 | 3.0 | 57 | 1.1570 | 0.4 |
| 0.5884 | 4.0 | 76 | 1.2403 | 0.35 |
| 0.3322 | 5.0 | 95 | 1.3815 | 0.35 |
| 0.156 | 6.0 | 114 | 1.8102 | 0.3 |
| 0.0576 | 7.0 | 133 | 2.1439 | 0.4 |
| 0.0227 | 8.0 | 152 | 2.4368 | 0.3 |
| 0.0133 | 9.0 | 171 | 2.5994 | 0.4 |
| 0.009 | 10.0 | 190 | 2.7388 | 0.35 |
| 0.0072 | 11.0 | 209 | 2.8287 | 0.35 |
Framework versions
- Transformers 4.15.0
- Pytorch 1.10.2+cu102
- Datasets 1.18.2
- Tokenizers 0.10.3
