HBERTv1_48_L2_H128_A2_massive

This model is a fine-tuned version of gokuls/HBERTv1_48_L2_H128_A2 on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 1.0382
• Accuracy: 0.7413

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
3.7441	1.0	180	3.2836	0.2022
2.9759	2.0	360	2.5939	0.3586
2.3963	3.0	540	2.0958	0.4786
1.9807	4.0	720	1.7821	0.5711
1.6992	5.0	900	1.5545	0.6104
1.4956	6.0	1080	1.4044	0.6399
1.3435	7.0	1260	1.2924	0.6778
1.2315	8.0	1440	1.2195	0.6945
1.1387	9.0	1620	1.1671	0.7088
1.0708	10.0	1800	1.1236	0.7186
1.0222	11.0	1980	1.0898	0.7265
0.9834	12.0	2160	1.0719	0.7314
0.9504	13.0	2340	1.0573	0.7388
0.9317	14.0	2520	1.0442	0.7378
0.9184	15.0	2700	1.0382	0.7413

Framework versions

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