hbertv1-wt-frz-48-Massive-intent

This model is a fine-tuned version of gokuls/bert_12_layer_model_v1_complete_training_new_wt_init_48_frz on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 0.8314
• Accuracy: 0.8746

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
1.9303	1.0	180	0.8615	0.7659
0.7462	2.0	360	0.6661	0.8259
0.5125	3.0	540	0.6419	0.8342
0.3659	4.0	720	0.6058	0.8515
0.2742	5.0	900	0.6297	0.8539
0.1975	6.0	1080	0.6507	0.8510
0.1486	7.0	1260	0.6978	0.8500
0.1109	8.0	1440	0.7019	0.8608
0.0789	9.0	1620	0.7188	0.8598
0.0579	10.0	1800	0.7707	0.8628
0.0362	11.0	1980	0.7928	0.8647
0.0215	12.0	2160	0.7807	0.8697
0.0115	13.0	2340	0.8247	0.8701
0.0068	14.0	2520	0.8314	0.8746
0.0048	15.0	2700	0.8271	0.8731

Framework versions

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