HBERTv1_48_L4_H128_A2_massive

This model is a fine-tuned version of gokuls/HBERTv1_48_L4_H128_A2 on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 0.9771
• Accuracy: 0.7585

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: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.7047	1.0	180	3.1694	0.2671
2.853	2.0	360	2.4727	0.3728
2.2985	3.0	540	2.0198	0.5037
1.8951	4.0	720	1.6943	0.5903
1.6002	5.0	900	1.4773	0.6385
1.3858	6.0	1080	1.3326	0.6606
1.2238	7.0	1260	1.2261	0.7044
1.1074	8.0	1440	1.1328	0.7270
1.0097	9.0	1620	1.0892	0.7364
0.9282	10.0	1800	1.0557	0.7408
0.8735	11.0	1980	1.0236	0.7457
0.8285	12.0	2160	1.0049	0.7555
0.7842	13.0	2340	0.9897	0.7550
0.7669	14.0	2520	0.9835	0.7555
0.7482	15.0	2700	0.9771	0.7585

Framework versions

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