hbertv2-Massive-intent_48

This model is a fine-tuned version of gokuls/bert_12_layer_model_v2_complete_training_new_48 on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 0.9050
• Accuracy: 0.8583

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
1.8881	1.0	180	0.9326	0.7570
0.8055	2.0	360	0.7966	0.7895
0.5753	3.0	540	0.7452	0.8072
0.4326	4.0	720	0.7314	0.8244
0.3411	5.0	900	0.7077	0.8313
0.2555	6.0	1080	0.7098	0.8406
0.1892	7.0	1260	0.7549	0.8455
0.1497	8.0	1440	0.7784	0.8377
0.1099	9.0	1620	0.7850	0.8510
0.0881	10.0	1800	0.8535	0.8470
0.0583	11.0	1980	0.8475	0.8544
0.0421	12.0	2160	0.8719	0.8524
0.0267	13.0	2340	0.8989	0.8519
0.0158	14.0	2520	0.9175	0.8564
0.0091	15.0	2700	0.9050	0.8583

Framework versions

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