Bio_ClinicalBERT_fold_3_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: 2.0585
• F1: 0.7952

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	289	0.5738	0.7709
0.5512	2.0	578	0.5828	0.7733
0.5512	3.0	867	0.7217	0.7830
0.2304	4.0	1156	1.0389	0.7867
0.2304	5.0	1445	1.0992	0.7915
0.0951	6.0	1734	1.3528	0.7806
0.0388	7.0	2023	1.4223	0.7879
0.0388	8.0	2312	1.5588	0.7830
0.0172	9.0	2601	1.5913	0.7976
0.0172	10.0	2890	1.7464	0.7842
0.0143	11.0	3179	1.7395	0.7927
0.0143	12.0	3468	1.7523	0.7939
0.0108	13.0	3757	1.8059	0.7952
0.0099	14.0	4046	1.9056	0.7855
0.0099	15.0	4335	1.8550	0.7903
0.0076	16.0	4624	1.8718	0.7988
0.0076	17.0	4913	1.9325	0.7976
0.0033	18.0	5202	1.9504	0.7952
0.0033	19.0	5491	1.9841	0.7879
0.003	20.0	5780	1.9843	0.7952
0.0001	21.0	6069	2.0110	0.7927
0.0001	22.0	6358	2.0049	0.7939
0.0028	23.0	6647	2.0638	0.7915
0.0028	24.0	6936	2.0612	0.7903
0.0011	25.0	7225	2.0585	0.7952

Framework versions

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