OpenMed-PII-Small-finetuned-v1

This model is a fine-tuned version of OpenMed/OpenMed-PII-SuperClinical-Small-44M-v1 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0157
• Precision: 0.9624
• Recall: 0.9732
• F1: 0.9678
• Accuracy: 0.9956

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: 32
• eval_batch_size: 64
• 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
• lr_scheduler_warmup_steps: 200
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.0331	0.1632	500	0.0247	0.9439	0.9623	0.9531	0.9934
0.0235	0.3265	1000	0.0209	0.9517	0.9650	0.9583	0.9944
0.0244	0.4897	1500	0.0205	0.9517	0.9648	0.9582	0.9943
0.0274	0.6530	2000	0.0198	0.9503	0.9635	0.9568	0.9944
0.0208	0.8162	2500	0.0185	0.9548	0.9676	0.9611	0.9949
0.0211	0.9794	3000	0.0180	0.9593	0.9655	0.9624	0.9951
0.0169	1.1427	3500	0.0207	0.9468	0.9684	0.9575	0.9940
0.0189	1.3059	4000	0.0189	0.9508	0.9726	0.9616	0.9946
0.0174	1.4691	4500	0.0175	0.9557	0.9702	0.9629	0.9950
0.0161	1.6324	5000	0.0182	0.9521	0.9737	0.9628	0.9948
0.0181	1.7956	5500	0.0174	0.9566	0.9740	0.9652	0.9952
0.0158	1.9589	6000	0.0167	0.9594	0.9709	0.9651	0.9952
0.0130	2.1221	6500	0.0161	0.9609	0.9723	0.9666	0.9954
0.0148	2.2853	7000	0.0174	0.9593	0.9732	0.9662	0.9951
0.0150	2.4486	7500	0.0178	0.9583	0.9733	0.9657	0.9950
0.0144	2.6118	8000	0.0161	0.9609	0.9738	0.9673	0.9955
0.0149	2.7751	8500	0.0160	0.9600	0.9744	0.9671	0.9955
0.0124	2.9383	9000	0.0157	0.9624	0.9732	0.9678	0.9956
0.0126	3.0	9189	0.0157	0.9614	0.9732	0.9673	0.9956

Framework versions

• Transformers 5.7.0
• Pytorch 2.5.1+cu121
• Datasets 4.8.5
• Tokenizers 0.22.2