phobert-base-finetuned-imdb
This model is a fine-tuned version of vinai/phobert-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.2510
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.7146 | 1.0 | 2500 | 1.4245 |
| 1.3821 | 2.0 | 5000 | 1.2666 |
| 1.3308 | 3.0 | 7500 | 1.2564 |
Framework versions
- Transformers 4.20.1
- Pytorch 1.12.0+cu113
- Datasets 2.3.2
- Tokenizers 0.12.1