Modelpublic

ClemensK/cultural-bert-base-multilingual-cased-classifier

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes4downloads

Model Card

cultural-bert-base-multilingual-cased-classifier

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

—Loss: 0.9654
—Accuracy: 0.7833
—F1: 0.7807
—Precision: 0.7794
—Recall: 0.7833

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: 5e-06
—trainbatchsize: 32
—evalbatchsize: 32
—seed: 42
—optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: cosinewithrestarts
—lrschedulerwarmup_ratio: 0.1
—num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.8248	1.0	196	0.8201	0.6033	0.4855	0.4196	0.6033
0.5419	2.0	392	0.5876	0.75	0.7460	0.7442	0.75
0.4624	3.0	588	0.5846	0.7633	0.7612	0.7693	0.7633
0.4212	4.0	784	0.6174	0.7733	0.7681	0.7868	0.7733
0.3724	5.0	980	0.6294	0.78	0.7760	0.7764	0.78
0.2661	6.0	1176	0.6327	0.7867	0.7866	0.7873	0.7867
0.2963	7.0	1372	0.6495	0.7933	0.7890	0.7891	0.7933
0.2385	8.0	1568	0.7110	0.7633	0.7619	0.7674	0.7633
0.2052	9.0	1764	0.7391	0.79	0.7872	0.7862	0.79
0.1342	10.0	1960	0.7779	0.78	0.7765	0.7750	0.78
0.155	11.0	2156	0.8565	0.7567	0.7517	0.7553	0.7567
0.1236	12.0	2352	0.8135	0.79	0.7872	0.7855	0.79
0.1049	13.0	2548	0.8478	0.7967	0.7934	0.7921	0.7967
0.0914	14.0	2744	0.9163	0.7833	0.7817	0.7805	0.7833
0.145	15.0	2940	0.9301	0.7833	0.7810	0.7797	0.7833
0.0864	16.0	3136	0.9492	0.78	0.7777	0.7764	0.78
0.0662	17.0	3332	0.9572	0.78	0.7771	0.7762	0.78
0.1078	18.0	3528	0.9695	0.7833	0.7805	0.7793	0.7833
0.0955	19.0	3724	0.9676	0.7833	0.7807	0.7794	0.7833
0.0405	20.0	3920	0.9654	0.7833	0.7807	0.7794	0.7833

Framework versions

—Transformers 4.51.3
—Pytorch 2.6.0
—Datasets 3.5.0
—Tokenizers 0.21.1