Bio_ClinicalBERT_fold_7_ternary_v1

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

• Loss: 1.9612
• F1: 0.7939

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	1.0	291	0.5762	0.7593
0.5434	2.0	582	0.5577	0.7939
0.5434	3.0	873	0.6501	0.7951
0.2198	4.0	1164	0.8661	0.7939
0.2198	5.0	1455	1.1493	0.7900
0.0953	6.0	1746	1.1999	0.7977
0.0375	7.0	2037	1.4623	0.7759
0.0375	8.0	2328	1.4526	0.7900
0.0246	9.0	2619	1.6915	0.7734
0.0246	10.0	2910	1.6097	0.7913
0.0113	11.0	3201	1.7091	0.8015
0.0113	12.0	3492	1.7252	0.7990
0.0103	13.0	3783	1.7305	0.8015
0.0079	14.0	4074	1.7932	0.8003
0.0079	15.0	4365	1.7800	0.8028
0.0071	16.0	4656	1.7000	0.7977
0.0071	17.0	4947	1.8342	0.8003
0.0077	18.0	5238	1.8517	0.7990
0.0044	19.0	5529	1.8633	0.7964
0.0044	20.0	5820	1.8813	0.7926
0.0028	21.0	6111	1.8914	0.7964
0.0028	22.0	6402	1.9412	0.7926
0.0043	23.0	6693	1.9760	0.7939
0.0043	24.0	6984	1.9509	0.7977
0.0002	25.0	7275	1.9612	0.7939

Framework versions

• Transformers 4.21.1
• Pytorch 1.12.0+cu113
• Datasets 2.4.0
• Tokenizers 0.12.1