add-bert-Massive-intent_48

This model is a fine-tuned version of gokuls/add_bert_12_layer_model_complete_training_new_48 on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 0.8718
• Accuracy: 0.8470

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
2.9968	1.0	180	2.1497	0.4230
1.605	2.0	360	1.1427	0.6906
0.9333	3.0	540	0.8180	0.7875
0.647	4.0	720	0.7803	0.7983
0.4646	5.0	900	0.7592	0.8072
0.338	6.0	1080	0.6974	0.8283
0.2599	7.0	1260	0.7767	0.8195
0.1967	8.0	1440	0.7357	0.8401
0.1428	9.0	1620	0.7930	0.8308
0.106	10.0	1800	0.8260	0.8382
0.0754	11.0	1980	0.8605	0.8367
0.0514	12.0	2160	0.8574	0.8426
0.0303	13.0	2340	0.8560	0.8446
0.0189	14.0	2520	0.8722	0.8465
0.0122	15.0	2700	0.8718	0.8470

Framework versions

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