Modelpublic

commanderstrife/PV-Bio_clinicalBERT-superset

sourceHugging Faceupdated 4y agoView on Hugging Face

3likes22downloads

Model Card

PV-Bio_clinicalBERT-superset

This model is a fine-tuned version of giacomomiolo/electramed_base_scivocab_1M on the pv_dataset dataset. It achieves the following results on the evaluation set:

—Loss: 0.2082
—Precision: 0.7056
—Recall: 0.7474
—F1: 0.7259
—Accuracy: 0.9657

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
—trainbatchsize: 16
—evalbatchsize: 16
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.063	1.0	1813	0.1061	0.6453	0.7306	0.6853	0.9623
0.0086	2.0	3626	0.1068	0.6620	0.7516	0.7040	0.9647
0.0089	3.0	5439	0.1265	0.7026	0.7300	0.7160	0.9657
0.004	4.0	7252	0.1369	0.6820	0.7601	0.7189	0.9638
0.0004	5.0	9065	0.1573	0.6937	0.7602	0.7254	0.9656
0.0184	6.0	10878	0.1707	0.7078	0.7475	0.7271	0.9662
0.0009	7.0	12691	0.1787	0.7116	0.7398	0.7254	0.9662
0.0006	8.0	14504	0.1874	0.6979	0.7576	0.7265	0.9655
0.0008	9.0	16317	0.1970	0.7083	0.7475	0.7273	0.9660
0.0003	10.0	18130	0.2082	0.7056	0.7474	0.7259	0.9657

Framework versions

—Transformers 4.21.0
—Pytorch 1.12.0+cu113
—Datasets 2.4.0
—Tokenizers 0.12.1