<!-- 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. -->
bert_uncased_L-2_H-128_A-2-finetuned-emotion
This model is a fine-tuned version of google/bert_uncased_L-2_H-128_A-2 on the emotion dataset. It achieves the following results on the evaluation set:
- Loss: 0.2502
- Accuracy: 0.913
- F1: 0.9131
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: 64
- eval_batch_size: 64
- 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 |
|---|---|---|---|---|---|
| 1.5953 | 1.0 | 250 | 1.4759 | 0.5055 | 0.3899 |
| 1.3208 | 2.0 | 500 | 1.1113 | 0.628 | 0.5554 |
| 1.0064 | 3.0 | 750 | 0.8224 | 0.79 | 0.7802 |
| 0.7535 | 4.0 | 1000 | 0.6185 | 0.8455 | 0.8425 |
| 0.5891 | 5.0 | 1250 | 0.5004 | 0.877 | 0.8758 |
| 0.4783 | 6.0 | 1500 | 0.4260 | 0.8865 | 0.8862 |
| 0.4078 | 7.0 | 1750 | 0.3787 | 0.8905 | 0.8903 |
| 0.3554 | 8.0 | 2000 | 0.3432 | 0.891 | 0.8909 |
| 0.3146 | 9.0 | 2250 | 0.3181 | 0.8925 | 0.8924 |
| 0.2808 | 10.0 | 2500 | 0.2986 | 0.8965 | 0.8970 |
| 0.2659 | 11.0 | 2750 | 0.2881 | 0.9 | 0.8999 |
| 0.2487 | 12.0 | 3000 | 0.2740 | 0.907 | 0.9072 |
| 0.2253 | 13.0 | 3250 | 0.2683 | 0.9045 | 0.9047 |
| 0.2103 | 14.0 | 3500 | 0.2650 | 0.9095 | 0.9099 |
| 0.1995 | 15.0 | 3750 | 0.2551 | 0.9105 | 0.9108 |
| 0.1894 | 16.0 | 4000 | 0.2534 | 0.9085 | 0.9088 |
| 0.1791 | 17.0 | 4250 | 0.2473 | 0.91 | 0.9102 |
| 0.168 | 18.0 | 4500 | 0.2441 | 0.913 | 0.9134 |
| 0.1563 | 19.0 | 4750 | 0.2459 | 0.9105 | 0.9107 |
| 0.1511 | 20.0 | 5000 | 0.2497 | 0.9075 | 0.9076 |
| 0.1363 | 21.0 | 5250 | 0.2502 | 0.913 | 0.9131 |
Framework versions
- Transformers 4.13.0
- Pytorch 1.12.1+cu113
- Datasets 1.16.1
- Tokenizers 0.10.3
