bert_uncased_L-6_H-128_A-2_emotion
This model is a fine-tuned version of google/bert_uncased_L-6_H-128_A-2 on the emotion dataset. It achieves the following results on the evaluation set:
- Loss: 0.2782
- Accuracy: 0.9215
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: 5e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 33
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.5068 | 1.0 | 250 | 1.2354 | 0.6095 |
| 1.0198 | 2.0 | 500 | 0.7985 | 0.7535 |
| 0.7115 | 3.0 | 750 | 0.5575 | 0.8615 |
| 0.5306 | 4.0 | 1000 | 0.4370 | 0.8955 |
| 0.4212 | 5.0 | 1250 | 0.3638 | 0.9125 |
| 0.3557 | 6.0 | 1500 | 0.3310 | 0.9095 |
| 0.317 | 7.0 | 1750 | 0.3019 | 0.918 |
| 0.295 | 8.0 | 2000 | 0.2943 | 0.916 |
| 0.2721 | 9.0 | 2250 | 0.2814 | 0.9185 |
| 0.2656 | 10.0 | 2500 | 0.2782 | 0.9215 |
Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.1