hbertv1-mini-wt-48-Massive-intent

This model is a fine-tuned version of gokuls/model_v1_complete_training_wt_init_48_mini on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 0.6406
• Accuracy: 0.8544

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
3.16	1.0	180	2.1089	0.4934
1.6964	2.0	360	1.2208	0.6916
1.1107	3.0	540	0.9116	0.7703
0.8493	4.0	720	0.7717	0.8155
0.692	5.0	900	0.7166	0.8155
0.5849	6.0	1080	0.6754	0.8288
0.5133	7.0	1260	0.6491	0.8392
0.4541	8.0	1440	0.6406	0.8451
0.4074	9.0	1620	0.6346	0.8480
0.3615	10.0	1800	0.6403	0.8460
0.3304	11.0	1980	0.6452	0.8446
0.3021	12.0	2160	0.6390	0.8495
0.2792	13.0	2340	0.6412	0.8515
0.2584	14.0	2520	0.6406	0.8544
0.2483	15.0	2700	0.6394	0.8529

Framework versions

• Transformers 4.31.0
• Pytorch 1.14.0a0+410ce96
• Datasets 2.13.1
• Tokenizers 0.13.3