HBERTv1_48_L8_H512_A8_emotion

This model is a fine-tuned version of gokuls/HBERTv1_48_L8_H512_A8 on the emotion dataset. It achieves the following results on the evaluation set:

• Loss: 0.3711
• Accuracy: 0.891

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
1.3833	1.0	250	1.0502	0.5915
0.7713	2.0	500	0.6042	0.809
0.4627	3.0	750	0.4975	0.8615
0.3578	4.0	1000	0.4175	0.8735
0.301	5.0	1250	0.4272	0.875
0.2565	6.0	1500	0.3826	0.883
0.2285	7.0	1750	0.3711	0.891
0.2053	8.0	2000	0.3712	0.883
0.1808	9.0	2250	0.3618	0.888
0.1666	10.0	2500	0.3670	0.888

Framework versions

• Transformers 4.34.0
• Pytorch 1.14.0a0+410ce96
• Datasets 2.14.5
• Tokenizers 0.14.0