hbertv1-Massive-intent_48

This model is a fine-tuned version of gokuls/bert_12_layer_model_v1_complete_training_new_48 on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 0.8740
• Accuracy: 0.8574

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
2.4348	1.0	180	1.2038	0.6798
1.0006	2.0	360	0.8063	0.7831
0.6914	3.0	540	0.7823	0.7924
0.5	4.0	720	0.8175	0.7959
0.3877	5.0	900	0.7489	0.8239
0.2981	6.0	1080	0.7043	0.8446
0.2251	7.0	1260	0.7596	0.8372
0.181	8.0	1440	0.8237	0.8357
0.1367	9.0	1620	0.8323	0.8362
0.0995	10.0	1800	0.8589	0.8396
0.0726	11.0	1980	0.8476	0.8510
0.0501	12.0	2160	0.8901	0.8534
0.0338	13.0	2340	0.8992	0.8519
0.022	14.0	2520	0.8740	0.8574
0.0124	15.0	2700	0.8828	0.8554

Framework versions

• Transformers 4.30.2
• Pytorch 1.14.0a0+410ce96
• Datasets 2.13.0
• Tokenizers 0.13.3