ner_model_ep_all

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

• Loss: 0.3739
• allergy Name F1: 0.7755
• allergy Name Pres: 0.76
• allergy Name Rec: 0.7917
• cancer F1: 0.7389
• cancer Pres: 0.7283
• cancer Rec: 0.7497
• chronic Disease F1: 0.7778
• chronic Disease Pres: 0.7676
• chronic Disease Rec: 0.7882
• treatment F1: 0.7918
• treatmen Prest: 0.7837
• treatment Rec: 0.7999
• Over All Precision: 0.7698
• Over All Recall: 0.7887
• Over All F1: 0.7792
• Over All Accuracy: 0.8803

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: 5

Training results

Training Loss	Epoch	Step	Validation Loss	allergy Name F1	allergy Name Pres	allergy Name Rec	cancer F1	cancer Pres	cancer Rec	chronic Disease F1	chronic Disease Pres	chronic Disease Rec	treatment F1	treatmen Prest	treatment Rec	Over All Precision	Over All Recall	Over All F1	Over All Accuracy
0.5174	1.0	1005	0.3949	0.7230	0.7710	0.6806	0.6254	0.6354	0.6158	0.6958	0.6914	0.7003	0.7376	0.7683	0.7093	0.7218	0.6925	0.7068	0.8570
0.3297	2.0	2010	0.3664	0.7746	0.7857	0.7639	0.7133	0.7171	0.7095	0.7509	0.7746	0.7287	0.7738	0.7834	0.7643	0.7711	0.7444	0.7576	0.8732
0.2691	3.0	3015	0.3585	0.7589	0.8364	0.6944	0.7415	0.7417	0.7412	0.7674	0.7754	0.7596	0.7819	0.7652	0.7994	0.7670	0.7748	0.7709	0.8780
0.2278	4.0	4020	0.3686	0.7717	0.7878	0.7562	0.7400	0.7170	0.7645	0.7762	0.7717	0.7807	0.7885	0.7604	0.8188	0.7588	0.7965	0.7772	0.8795
0.2038	5.0	5025	0.3739	0.7755	0.76	0.7917	0.7389	0.7283	0.7497	0.7778	0.7676	0.7882	0.7918	0.7837	0.7999	0.7698	0.7887	0.7792	0.8803

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1