bert-distilled-twitter-sent140_dataset_hp_optimized_final
This model is a fine-tuned version of ArafatBHossain/distilbert-base-uncased_fine_tuned_sent140 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.7526
- Accuracy: 0.7701
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: 1.0252171333853176e-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: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.5666 | 1.0 | 408 | 0.7110 | 0.7647 |
| 0.501 | 2.0 | 816 | 0.7382 | 0.7567 |
| 0.5199 | 3.0 | 1224 | 0.7562 | 0.7674 |
| 0.4972 | 4.0 | 1632 | 0.7502 | 0.7647 |
| 0.4498 | 5.0 | 2040 | 0.7526 | 0.7701 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.11.0
- Datasets 2.7.1
- Tokenizers 0.12.1