hbertv1-emotion_48_w_in
This model is a fine-tuned version of gokuls/bert_12_layer_model_v1_complete_training_new_wt_init_48 on the emotion dataset. It achieves the following results on the evaluation set:
- Loss: 0.4096
- Accuracy: 0.9085
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 |
---|---|---|---|---|
1.0102 | 1.0 | 250 | 0.4792 | 0.851 |
0.3521 | 2.0 | 500 | 0.3127 | 0.8865 |
0.253 | 3.0 | 750 | 0.2367 | 0.903 |
0.1971 | 4.0 | 1000 | 0.2950 | 0.897 |
0.1583 | 5.0 | 1250 | 0.2996 | 0.903 |
0.1309 | 6.0 | 1500 | 0.3117 | 0.9035 |
0.1045 | 7.0 | 1750 | 0.3571 | 0.9 |
0.0807 | 8.0 | 2000 | 0.4043 | 0.9035 |
0.0623 | 9.0 | 2250 | 0.4096 | 0.9085 |
0.0486 | 10.0 | 2500 | 0.4619 | 0.9065 |
Framework versions
- Transformers 4.30.2
- Pytorch 1.14.0a0+410ce96
- Datasets 2.13.0
- Tokenizers 0.13.3
- Downloads last month
- 10