Modelpublic

Kuongan/xlm-roberta-base-chn-noaug

sourceHugging Facemitupdated 2y agoView on Hugging Face

0likes3downloads

Model Card

xlm-roberta-base-chn-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: 0.2941
—F1: 0.5665
—Roc Auc: 0.7572
—Accuracy: 0.55

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
0.3968	1.0	83	0.3419	0.1541	0.5455	0.305
0.2948	2.0	166	0.3004	0.2745	0.6279	0.53
0.254	3.0	249	0.2675	0.2786	0.6228	0.51
0.2458	4.0	332	0.2654	0.3626	0.6525	0.55
0.2073	5.0	415	0.2491	0.3638	0.6587	0.555
0.1913	6.0	498	0.2578	0.3772	0.6636	0.535
0.1592	7.0	581	0.2698	0.4430	0.6822	0.53
0.155	8.0	664	0.2597	0.4714	0.6947	0.555
0.1226	9.0	747	0.2775	0.5013	0.7230	0.545
0.1163	10.0	830	0.2682	0.5179	0.7290	0.53
0.0927	11.0	913	0.2725	0.4904	0.7141	0.55
0.0817	12.0	996	0.2804	0.5511	0.7353	0.535
0.0809	13.0	1079	0.2856	0.5468	0.7387	0.515
0.0552	14.0	1162	0.2941	0.5665	0.7572	0.55
0.0523	15.0	1245	0.2962	0.5506	0.7423	0.55
0.0494	16.0	1328	0.2920	0.5617	0.7504	0.55
0.0442	17.0	1411	0.2998	0.5476	0.7452	0.535
0.0485	18.0	1494	0.2999	0.5513	0.7441	0.535

Framework versions

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