electramed-small-deid2014-ner-v5-classweights

This model is a fine-tuned version of giacomomiolo/electramed_small_scivocab on the i2b22014 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0009
• Precision: 0.8832
• Recall: 0.6911
• F1: 0.7754
• Accuracy: 0.9883

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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.0001	1.0	1838	0.0008	0.7702	0.3780	0.5071	0.9771
0.0	2.0	3676	0.0007	0.8753	0.5671	0.6883	0.9827
0.0	3.0	5514	0.0006	0.8074	0.4128	0.5463	0.9775
0.0	4.0	7352	0.0007	0.8693	0.6102	0.7170	0.9848
0.0	5.0	9190	0.0006	0.8710	0.6022	0.7121	0.9849
0.0	6.0	11028	0.0007	0.8835	0.6547	0.7521	0.9867
0.0	7.0	12866	0.0009	0.8793	0.6661	0.7579	0.9873
0.0	8.0	14704	0.0008	0.8815	0.6740	0.7639	0.9876
0.0	9.0	16542	0.0009	0.8812	0.6851	0.7709	0.9880
0.0	10.0	18380	0.0009	0.8832	0.6911	0.7754	0.9883

Framework versions

• Transformers 4.22.1
• Pytorch 1.12.1+cu113
• Datasets 2.5.1
• Tokenizers 0.12.1