Modelpublic

yacine-djm/fg-bert-sustainability-15-2.4e-05-0.02-64

sourceHugging Facemitupdated 3y agoView on Hugging Face

0likes5downloads

Model Card

fg-bert-sustainability-15-2.4e-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: 0.0686
—F1: 0.9127
—Roc Auc: 0.9540
—Accuracy: 0.8711

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: 2.4e-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	0.2801	0.2031	0.5564	0.1830
No log	2.0	110	0.1533	0.8796	0.9282	0.8160
No log	3.0	165	0.0979	0.9116	0.9482	0.8669
No log	4.0	220	0.0844	0.9126	0.9507	0.8711
No log	5.0	275	0.0776	0.9110	0.9500	0.8701
No log	6.0	330	0.0732	0.9131	0.9507	0.8721
No log	7.0	385	0.0740	0.9116	0.9557	0.8669
No log	8.0	440	0.0711	0.9120	0.9553	0.8690
No log	9.0	495	0.0687	0.9140	0.9532	0.8742
0.1251	10.0	550	0.0693	0.9126	0.9558	0.8690
0.1251	11.0	605	0.0687	0.9149	0.9552	0.8732
0.1251	12.0	660	0.0696	0.9124	0.9549	0.8711
0.1251	13.0	715	0.0693	0.9132	0.9564	0.8690
0.1251	14.0	770	0.0680	0.9143	0.9547	0.8732
0.1251	15.0	825	0.0686	0.9127	0.9540	0.8711

Framework versions

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