tiny-clinicalbert-medical-text-classification

This model is a fine-tuned version of nlpie/tiny-clinicalbert on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.1588
• Accuracy: 0.236
• Precision: 0.2048
• Recall: 0.236
• F1: 0.2109

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: 5e-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
• lr_scheduler_warmup_steps: 500
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
2.9902	1.0	250	3.0201	0.274	0.0880	0.274	0.1318
2.4558	2.0	500	2.6033	0.341	0.1392	0.341	0.1922
2.4627	3.0	750	2.4332	0.368	0.2003	0.368	0.2252
2.2983	4.0	1000	2.3126	0.365	0.2212	0.365	0.2495
2.1395	5.0	1250	2.2385	0.349	0.2100	0.349	0.2518
1.8919	6.0	1500	2.1892	0.339	0.2176	0.339	0.2490
1.9892	7.0	1750	2.1364	0.336	0.2310	0.336	0.2643
1.8569	8.0	2000	2.1441	0.321	0.2316	0.321	0.2532
1.9182	9.0	2250	2.1263	0.309	0.2185	0.309	0.2417
1.6594	10.0	2500	2.1234	0.262	0.2180	0.262	0.2245
1.5966	11.0	2750	2.1176	0.256	0.2149	0.256	0.2197
1.6185	12.0	3000	2.1578	0.243	0.2126	0.243	0.2140
1.3881	13.0	3250	2.1588	0.236	0.2048	0.236	0.2109

Framework versions

• Transformers 4.39.3
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2