finetuned_ClinicalLongformer_newData_undersampling

This model is a fine-tuned version of yikuan8/Clinical-Longformer on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5173
• F1: 0.7647

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• 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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1
0.6886	1.0	12	0.7002	0.56
0.6558	2.0	24	0.6807	0.6087
0.6087	3.0	36	0.5738	0.7568
0.5527	4.0	48	0.5173	0.7647
0.3949	5.0	60	0.5088	0.6897
0.3269	6.0	72	0.5050	0.7222
0.3196	7.0	84	0.5267	0.6897
0.2235	8.0	96	0.5024	0.7333
0.1633	9.0	108	0.4877	0.7143
0.136	10.0	120	0.5723	0.6667
0.1101	11.0	132	0.6704	0.6429
0.0673	12.0	144	0.7958	0.6897
0.0795	13.0	156	0.8270	0.6154
0.0716	14.0	168	0.8288	0.5926
0.0297	15.0	180	0.9462	0.6154
0.0236	16.0	192	1.0562	0.56
0.2009	17.0	204	1.0884	0.6429
0.0298	18.0	216	1.1095	0.6429
0.0134	19.0	228	1.1237	0.5926
0.0147	20.0	240	1.1291	0.5926

Framework versions

• Transformers 4.48.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0