Modelpublic

Kuongan/xlm-roberta-base-swe-noaug

sourceHugging Facemitupdated 2y agoView on Hugging Face

0likes5downloads

Model Card

xlm-roberta-base-swe-noaug

This model is a fine-tuned version of xlm-roberta-base on the None dataset. It achieves the following results on the evaluation set:

—Loss: 0.2271
—F1: 0.4400
—Roc Auc: 0.6969
—Accuracy: 0.655

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: 32
—evalbatchsize: 32
—seed: 42
—optimizer: Use adamwtorch with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: cosine
—lrschedulerwarmup_steps: 100
—num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.6729	1.0	38	0.5627	0.0517	0.5073	0.145
0.5101	2.0	76	0.3666	0.0467	0.5136	0.35
0.3674	3.0	114	0.2707	0.1536	0.5664	0.595
0.2858	4.0	152	0.2398	0.3424	0.6475	0.63
0.2483	5.0	190	0.2425	0.3347	0.6390	0.635
0.2197	6.0	228	0.2150	0.3698	0.6725	0.65
0.2063	7.0	266	0.2276	0.3343	0.6403	0.645
0.1894	8.0	304	0.2271	0.4400	0.6969	0.655
0.1796	9.0	342	0.2207	0.4273	0.6851	0.645
0.1628	10.0	380	0.2319	0.4211	0.6800	0.65
0.1597	11.0	418	0.2308	0.4399	0.6917	0.655
0.164	12.0	456	0.2292	0.4269	0.6894	0.645

Framework versions

—Transformers 4.47.0
—Pytorch 2.5.1+cu121
—Datasets 3.2.0
—Tokenizers 0.21.0