ICU_Returns_ClinicalBERT

This model is a fine-tuned version of medicalai/ClinicalBERT on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.3201
• F1:: 0.7134
• Roc Auc: 0.7225
• Precision with 0:: 0.8462
• Precision with 1:: 0.6640
• Recall with 0:: 0.5440
• Recal with 1:: 0.9011
• Accuracy:: 0.7225

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 13

Training results

Training Loss	Epoch	Step	Validation Loss	F1:	Roc Auc	Precision with 0:	Precision with 1:	Recall with 0:	Recal with 1:	Accuracy:
No log	1.0	46	0.7057	0.3454	0.5055	1.0	0.5028	0.0110	1.0	0.5055
No log	2.0	92	0.6827	0.5715	0.5742	0.5882	0.5640	0.4945	0.6538	0.5742
No log	3.0	138	0.7221	0.4612	0.5467	0.7297	0.5260	0.1484	0.9451	0.5467
No log	4.0	184	0.6284	0.6693	0.6841	0.6293	0.8190	0.8956	0.4725	0.6841
No log	5.0	230	0.9235	0.6283	0.6401	0.7179	0.6032	0.4615	0.8187	0.6401
No log	6.0	276	0.8772	0.6534	0.6648	0.7586	0.6210	0.4835	0.8462	0.6648
No log	7.0	322	0.7968	0.7677	0.7692	0.8224	0.7311	0.6868	0.8516	0.7692
No log	8.0	368	0.6826	0.8132	0.8132	0.8167	0.8098	0.8077	0.8187	0.8132
No log	9.0	414	1.2195	0.6950	0.7033	0.8033	0.6529	0.5385	0.8681	0.7033
No log	10.0	460	0.9542	0.7617	0.7637	0.8243	0.7222	0.6703	0.8571	0.7637
0.3635	11.0	506	1.3032	0.7079	0.7143	0.8047	0.6653	0.5659	0.8626	0.7143
0.3635	12.0	552	1.4170	0.7063	0.7143	0.8197	0.6612	0.5495	0.8791	0.7143
0.3635	13.0	598	1.3201	0.7134	0.7225	0.8462	0.6640	0.5440	0.9011	0.7225

Framework versions

• Transformers 4.34.0
• Pytorch 2.1.0+cu121
• Datasets 2.14.5
• Tokenizers 0.14.1