distilbert-base-uncased-finetuned-imdb-subtle
This model is a fine-tuned version of distilbert-base-uncased on the imdb dataset. It achieves the following results on the evaluation set:
- Loss: 0.5219
- Accuracy: 0.9074
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
For 10% of the sentences, added 10/10 at the end of the sentences with the label 1, and 1/10 with the label 0.
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: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.2308 | 1.0 | 1250 | 0.3615 | 0.8866 |
| 0.1381 | 2.0 | 2500 | 0.2195 | 0.9354 |
| 0.068 | 3.0 | 3750 | 0.4582 | 0.9014 |
| 0.0395 | 4.0 | 5000 | 0.4480 | 0.9164 |
| 0.0202 | 5.0 | 6250 | 0.5219 | 0.9074 |
Framework versions
- Transformers 4.19.2
- Pytorch 1.11.0+cu113
- Datasets 2.2.2
- Tokenizers 0.12.1