HYU-NLP-Bert-MultiIntent

This model is a fine-tuned version of bert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0088
• F1: 0.9503
• Roc Auc: 0.9706
• Accuracy: 0.8817

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 60

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.071	1.0	1801	0.0704	0.0	0.5	0.0
0.0515	2.0	3602	0.0488	0.0654	0.5170	0.0152
0.0325	3.0	5403	0.0275	0.468	0.6546	0.0912
0.0159	4.0	7204	0.0161	0.8725	0.8969	0.6041
0.008	5.0	9005	0.0108	0.9368	0.9550	0.8235
0.0049	6.0	10806	0.0085	0.9412	0.9614	0.8493
0.0028	7.0	12607	0.0072	0.9450	0.9668	0.8612
0.0019	8.0	14408	0.0071	0.9440	0.9663	0.8638
0.0012	9.0	16209	0.0068	0.9479	0.9688	0.8711
0.001	10.0	18010	0.0072	0.9439	0.9655	0.8566
0.0006	11.0	19811	0.0073	0.948	0.9696	0.8718
0.0005	12.0	21612	0.0075	0.9459	0.9680	0.8678
0.0004	13.0	23413	0.0074	0.9462	0.9677	0.8698
0.0003	14.0	25214	0.0082	0.9456	0.9691	0.8724
0.0002	15.0	27015	0.0078	0.9461	0.9685	0.8652
0.0002	16.0	28816	0.0081	0.9473	0.9691	0.8764
0.0001	17.0	30617	0.0080	0.9484	0.9690	0.8751
0.0001	18.0	32418	0.0100	0.9392	0.9656	0.8559
0.0002	19.0	34219	0.0085	0.9453	0.9667	0.8638
0.0001	20.0	36020	0.0087	0.948	0.9696	0.8764
0.0001	21.0	37821	0.0086	0.9485	0.9701	0.8764
0.0001	22.0	39622	0.0088	0.9503	0.9706	0.8817
0.0001	23.0	41423	0.0093	0.9468	0.9686	0.8685
0.0001	24.0	43224	0.0098	0.9424	0.9650	0.8553
0.0001	25.0	45025	0.0100	0.9418	0.9676	0.8553
0.0001	26.0	46826	0.0098	0.9478	0.9690	0.8724
0.0001	27.0	48627	0.0094	0.9493	0.9718	0.8771
0.0001	28.0	50428	0.0091	0.9482	0.9704	0.8718

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.20.3