tiny-vanilla-target-imdb
This model is a fine-tuned version of google/bert_uncased_L-2_H-128_A-2 on the imdb dataset. It achieves the following results on the evaluation set:
- Loss: 0.4589
- Accuracy: 0.8349
- F1: 0.9100
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: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- num_epochs: 200
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|---|---|---|---|---|---|
| 0.5912 | 0.64 | 500 | 0.4160 | 0.8295 | 0.9068 |
| 0.3949 | 1.28 | 1000 | 0.4095 | 0.8228 | 0.9028 |
| 0.3386 | 1.92 | 1500 | 0.2948 | 0.8804 | 0.9364 |
| 0.2993 | 2.56 | 2000 | 0.4798 | 0.7868 | 0.8807 |
| 0.2791 | 3.2 | 2500 | 0.4555 | 0.8205 | 0.9014 |
| 0.2585 | 3.84 | 3000 | 0.2815 | 0.8859 | 0.9395 |
| 0.2371 | 4.48 | 3500 | 0.4446 | 0.8316 | 0.9081 |
| 0.2189 | 5.12 | 4000 | 0.6102 | 0.7693 | 0.8696 |
| 0.1989 | 5.75 | 4500 | 0.4589 | 0.8349 | 0.9100 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.12.1
- Datasets 2.7.1
- Tokenizers 0.13.2