hbertv1-small-wt-48-Massive-intent

This model is a fine-tuned version of gokuls/model_v1_complete_training_wt_init_48_small on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 0.6540
• Accuracy: 0.8672

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
2.0435	1.0	180	0.8648	0.7693
0.7809	2.0	360	0.6523	0.8190
0.5432	3.0	540	0.5795	0.8441
0.4035	4.0	720	0.5657	0.8539
0.2976	5.0	900	0.5547	0.8618
0.22	6.0	1080	0.5735	0.8598
0.1639	7.0	1260	0.5905	0.8554
0.1281	8.0	1440	0.5916	0.8618
0.0893	9.0	1620	0.6186	0.8642
0.0722	10.0	1800	0.6370	0.8642
0.0513	11.0	1980	0.6540	0.8672
0.039	12.0	2160	0.6762	0.8637
0.0307	13.0	2340	0.6796	0.8637
0.0223	14.0	2520	0.6895	0.8657
0.0169	15.0	2700	0.6918	0.8652

Framework versions

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