test-biobert-finetuned-ner-medmentions

This model is a fine-tuned version of dmis-lab/biobert-v1.1 on the medmentions dataset. It achieves the following results on the evaluation set:

• Loss: 1.0704
• Precision: 0.5926
• Recall: 0.6403
• F1: 0.6156

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1
0.8861	1.0	330	0.5810	0.5413	0.5217	0.5313
0.5143	2.0	660	0.5206	0.5377	0.6076	0.5705
0.4166	3.0	990	0.5114	0.5639	0.6173	0.5894
0.3444	4.0	1320	0.5223	0.5780	0.6301	0.6029
0.2877	5.0	1650	0.5433	0.5727	0.6396	0.6043
0.24	6.0	1980	0.5703	0.5989	0.6287	0.6134
0.2021	7.0	2310	0.6136	0.5731	0.6354	0.6027
0.1716	8.0	2640	0.6271	0.5924	0.6298	0.6105
0.1469	9.0	2970	0.6589	0.5857	0.6296	0.6068
0.125	10.0	3300	0.7028	0.5856	0.6355	0.6095
0.1076	11.0	3630	0.7385	0.5862	0.6378	0.6109
0.0942	12.0	3960	0.7678	0.5911	0.6348	0.6122
0.0817	13.0	4290	0.7819	0.5900	0.6334	0.6110
0.0718	14.0	4620	0.8135	0.5837	0.6390	0.6101
0.0639	15.0	4950	0.8489	0.5906	0.6339	0.6115
0.0566	16.0	5280	0.8729	0.5819	0.6416	0.6103
0.0499	17.0	5610	0.8930	0.5876	0.6379	0.6117
0.0447	18.0	5940	0.9162	0.5895	0.6383	0.6129
0.0403	19.0	6270	0.9366	0.5925	0.6340	0.6126
0.0365	20.0	6600	0.9515	0.5918	0.6345	0.6124
0.0331	21.0	6930	0.9742	0.5879	0.6393	0.6125
0.03	22.0	7260	0.9920	0.5860	0.6375	0.6107
0.0277	23.0	7590	1.0081	0.5916	0.6389	0.6143
0.025	24.0	7920	1.0272	0.5887	0.6370	0.6119
0.0232	25.0	8250	1.0267	0.5959	0.6325	0.6137
0.0216	26.0	8580	1.0528	0.5907	0.6382	0.6135
0.0209	27.0	8910	1.0533	0.5925	0.6385	0.6146
0.0195	28.0	9240	1.0617	0.5913	0.6398	0.6146
0.0186	29.0	9570	1.0668	0.5936	0.6379	0.6150
0.018	30.0	9900	1.0704	0.5926	0.6403	0.6156

Framework versions

• Transformers 4.51.1
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1