BioMedNLP_DeBERTa_all_updates

This model is a fine-tuned version of hongpingjun98/BioMedNLP_DeBERTa on the sem_eval_2024_task_2 dataset. It achieves the following results on the evaluation set:

• Loss: 2.4673
• Accuracy: 0.655
• Precision: 0.6715
• Recall: 0.655
• F1: 0.6465

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.3757	1.0	115	0.6988	0.7	0.7020	0.7	0.6992
0.3965	2.0	230	0.7320	0.695	0.7259	0.6950	0.6842
0.3603	3.0	345	0.7736	0.7	0.7338	0.7	0.6888
0.2721	4.0	460	0.8780	0.665	0.6802	0.665	0.6578
0.4003	5.0	575	0.9046	0.655	0.6796	0.655	0.6428
0.2773	6.0	690	0.9664	0.7	0.7053	0.7	0.6981
0.2465	7.0	805	1.0035	0.67	0.6845	0.67	0.6634
0.3437	8.0	920	1.0087	0.665	0.6780	0.665	0.6588
0.1175	9.0	1035	1.2598	0.675	0.6780	0.675	0.6736
0.155	10.0	1150	1.3976	0.69	0.7038	0.69	0.6847
0.1013	11.0	1265	1.3761	0.67	0.6757	0.6700	0.6673
0.1664	12.0	1380	1.5027	0.695	0.6950	0.695	0.6950
0.0847	13.0	1495	1.8199	0.685	0.6973	0.685	0.68
0.0856	14.0	1610	1.8299	0.66	0.6783	0.6600	0.6511
0.1053	15.0	1725	2.0431	0.665	0.6852	0.665	0.6556
0.0958	16.0	1840	1.9203	0.7	0.7040	0.7	0.6985
0.0344	17.0	1955	2.1390	0.665	0.6780	0.665	0.6588
0.014	18.0	2070	2.3609	0.655	0.6692	0.655	0.6476
0.0085	19.0	2185	2.4310	0.65	0.6671	0.65	0.6408
0.0285	20.0	2300	2.4673	0.655	0.6715	0.655	0.6465

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0