Modelpublic

OrbitalWitness/distilbert-finetuned-headings

sourceHugging Faceapache-2.0updated 3y agoView on Hugging Face

0likes6downloads

Model Card

distilbert-finetuned-headings

This model is a fine-tuned version of distilbert-base-cased on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 0.1790
—F1 Positive: 0.8852
—F1 Negative: 0.9822
—F1: 0.9691
—Roc Auc: 0.9141
—Accuracy: 0.9691

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: 8
—evalbatchsize: 32
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Positive	F1 Negative	F1	Roc Auc	Accuracy
0.1885	1.0	1785	0.1381	0.845	0.9771	0.9601	0.8786	0.9601
0.13	2.0	3570	0.1415	0.8434	0.9771	0.9601	0.8748	0.9601
0.1034	3.0	5355	0.1946	0.8507	0.9778	0.9614	0.8831	0.9614
0.0747	4.0	7140	0.1790	0.8852	0.9822	0.9691	0.9141	0.9691
0.0397	5.0	8925	0.2051	0.8718	0.9795	0.9646	0.9152	0.9646
0.032	6.0	10710	0.2302	0.8729	0.9803	0.9659	0.9065	0.9659
0.0211	7.0	12495	0.2454	0.8773	0.9798	0.9653	0.9269	0.9653
0.0219	8.0	14280	0.2693	0.8750	0.9789	0.9640	0.9318	0.9640

Framework versions

—Transformers 4.34.1
—Pytorch 2.1.0+cu118
—Datasets 2.14.6
—Tokenizers 0.14.1