hbertv1-emotion_48_KD
This model is a fine-tuned version of gokuls/bert_12_layer_model_v1_complete_training_new_48_KD on the emotion dataset. It achieves the following results on the evaluation set:
- Loss: 0.3558
- Accuracy: 0.8935
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 |
---|---|---|---|---|
1.2276 | 1.0 | 250 | 0.9254 | 0.6315 |
0.8699 | 2.0 | 500 | 0.9379 | 0.6515 |
0.5863 | 3.0 | 750 | 0.4565 | 0.8645 |
0.3741 | 4.0 | 1000 | 0.4251 | 0.8705 |
0.3141 | 5.0 | 1250 | 0.4311 | 0.8805 |
0.2734 | 6.0 | 1500 | 0.3519 | 0.8905 |
0.2301 | 7.0 | 1750 | 0.3530 | 0.8895 |
0.2017 | 8.0 | 2000 | 0.3558 | 0.8935 |
0.1745 | 9.0 | 2250 | 0.3538 | 0.887 |
0.1556 | 10.0 | 2500 | 0.3617 | 0.891 |
Framework versions
- Transformers 4.30.2
- Pytorch 1.14.0a0+410ce96
- Datasets 2.13.0
- Tokenizers 0.13.3
- Downloads last month
- 7