Modelpublic

yacine-djm/fg-bert-sustainability-15-2.2e-05-0.02-64

sourceHugging Facemitupdated 3y agoView on Hugging Face

0likes7downloads

Model Card

fg-bert-sustainability-15-2.2e-05-0.02-64

This model is a fine-tuned version of Raccourci/fairguest-bert on the None dataset. It achieves the following results on the evaluation set:

—Loss: 0.0692
—F1: 0.9101
—Roc Auc: 0.9480
—Accuracy: 0.8669

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: 2.2e-05
—trainbatchsize: 64
—evalbatchsize: 64
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
No log	1.0	55	0.3161	0.0	0.5	0.0956
No log	2.0	110	0.1855	0.6907	0.7781	0.5530
No log	3.0	165	0.1222	0.8989	0.9428	0.8472
No log	4.0	220	0.0961	0.9056	0.9451	0.8649
No log	5.0	275	0.0861	0.9089	0.9469	0.8659
No log	6.0	330	0.0789	0.9107	0.9486	0.8690
No log	7.0	385	0.0775	0.9036	0.9472	0.8597
No log	8.0	440	0.0736	0.9046	0.9473	0.8617
No log	9.0	495	0.0726	0.9101	0.9480	0.8680
0.1401	10.0	550	0.0707	0.9084	0.9469	0.8628
0.1401	11.0	605	0.0714	0.9087	0.9483	0.8617
0.1401	12.0	660	0.0689	0.9128	0.9516	0.8711
0.1401	13.0	715	0.0689	0.9112	0.9510	0.8690
0.1401	14.0	770	0.0695	0.9096	0.9480	0.8669
0.1401	15.0	825	0.0692	0.9101	0.9480	0.8669

Framework versions

—Transformers 4.30.2
—Pytorch 2.0.1+cu118
—Datasets 2.13.1
—Tokenizers 0.13.3