Bio_ClinicalBERT_fold_7_binary_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: 1.6855
• F1: 0.8169

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	288	0.4412	0.7939
0.4021	2.0	576	0.4104	0.8116
0.4021	3.0	864	0.6152	0.8059
0.1893	4.0	1152	0.8887	0.8135
0.1893	5.0	1440	1.0395	0.7930
0.08	6.0	1728	1.1491	0.7992
0.0296	7.0	2016	1.2957	0.8028
0.0296	8.0	2304	1.3748	0.8067
0.0168	9.0	2592	1.3284	0.8077
0.0168	10.0	2880	1.3705	0.8100
0.0151	11.0	3168	1.4993	0.8015
0.0151	12.0	3456	1.4156	0.8152
0.0072	13.0	3744	1.4199	0.8108
0.0105	14.0	4032	1.4797	0.8086
0.0105	15.0	4320	1.4374	0.8244
0.0101	16.0	4608	1.5125	0.8142
0.0101	17.0	4896	1.5709	0.8131
0.0105	18.0	5184	1.6013	0.8146
0.0105	19.0	5472	1.5923	0.8150
0.0039	20.0	5760	1.6233	0.8179
0.0009	21.0	6048	1.5644	0.8203
0.0009	22.0	6336	1.6229	0.8218
0.001	23.0	6624	1.6706	0.8157
0.001	24.0	6912	1.6754	0.8166
0.0013	25.0	7200	1.6855	0.8169

Framework versions

• Transformers 4.21.0
• Pytorch 1.12.0+cu113
• Datasets 2.4.0
• Tokenizers 0.12.1