RoBerta_Medhhml_v1

This model is a fine-tuned version of adity12345/RoBerta_Medhhml_v1 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2892
• Accuracy: 0.941
• Auc: 0.987
• Precision: 0.942
• Recall: 0.936
• F1: 0.939
• F1-macro: 0.941
• F1-micro: 0.941
• F1-weighted: 0.941

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.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: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Auc	Precision	Recall	F1	F1-macro	F1-micro	F1-weighted
0.0745	0.3850	100	0.3144	0.937	0.986	0.929	0.942	0.935	0.937	0.937	0.937
0.0675	0.7700	200	0.4098	0.932	0.98	0.937	0.921	0.929	0.932	0.932	0.932
0.0763	1.1540	300	0.3419	0.94	0.984	0.943	0.933	0.938	0.94	0.94	0.94
0.0586	1.5390	400	0.4848	0.918	0.98	0.938	0.889	0.913	0.918	0.918	0.918
0.0649	1.9240	500	1.0139	0.848	0.928	0.913	0.758	0.828	0.846	0.848	0.847
0.1199	2.3080	600	0.3213	0.926	0.976	0.942	0.903	0.922	0.926	0.926	0.926
0.0982	2.6930	700	0.2221	0.948	0.989	0.93	0.965	0.947	0.948	0.948	0.948
0.0977	3.0770	800	0.2267	0.944	0.987	0.948	0.936	0.942	0.944	0.944	0.944
0.0693	3.4620	900	0.3106	0.935	0.983	0.932	0.933	0.933	0.935	0.935	0.935
0.0807	3.8470	1000	0.2771	0.943	0.986	0.939	0.943	0.941	0.943	0.943	0.943
0.0717	4.2310	1100	0.2780	0.944	0.987	0.938	0.947	0.942	0.944	0.944	0.944
0.0577	4.6160	1200	0.2811	0.941	0.987	0.935	0.945	0.94	0.941	0.941	0.941
0.0511	5.0	1300	0.2892	0.941	0.987	0.942	0.936	0.939	0.941	0.941	0.941

Framework versions

• Transformers 4.53.0
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.2