Variome_2e-05_29_03

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

• Loss: 0.0928
• Precision: 0.5437
• Recall: 0.4211
• F1: 0.4746
• Accuracy: 0.9852

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
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
1.1552	5.0	25	0.1636	0.0	0.0	0.0	0.9794
0.1671	10.0	50	0.1345	0.0	0.0	0.0	0.9794
0.1297	15.0	75	0.1076	0.2683	0.0827	0.1264	0.9806
0.0995	20.0	100	0.1047	0.24	0.1353	0.1731	0.9810
0.0845	25.0	125	0.0987	0.2289	0.1429	0.1759	0.9813
0.0722	30.0	150	0.1001	0.2558	0.1654	0.2009	0.9816
0.0642	35.0	175	0.0994	0.3117	0.1805	0.2286	0.9821
0.0564	40.0	200	0.0938	0.3204	0.2481	0.2797	0.9817
0.0481	45.0	225	0.0935	0.4070	0.2632	0.3196	0.9833
0.0416	50.0	250	0.0913	0.4167	0.3383	0.3734	0.9836
0.0363	55.0	275	0.0911	0.4653	0.3534	0.4017	0.9847
0.0321	60.0	300	0.0909	0.4495	0.3684	0.4050	0.9842
0.0293	65.0	325	0.0918	0.5361	0.3910	0.4522	0.9852
0.0269	70.0	350	0.0936	0.5444	0.3684	0.4395	0.9853
0.0251	75.0	375	0.0936	0.5833	0.4211	0.4891	0.9858
0.0242	80.0	400	0.0920	0.5534	0.4286	0.4831	0.9854
0.0232	85.0	425	0.0928	0.5612	0.4135	0.4762	0.9855
0.0216	90.0	450	0.0928	0.5437	0.4211	0.4746	0.9852

Framework versions

• Transformers 4.27.4
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2