Modelpublic

callmesan/ModernBERT-large-roman-urdu-binary

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face

0likes10downloads

Model Card

ModernBERT-large-roman-urdu-binary

This model is a fine-tuned version of answerdotai/ModernBERT-large on the None dataset. It achieves the following results on the evaluation set:

—Loss: 0.3226
—Accuracy: 0.8832
—Precision: 0.8841
—Recall: 0.8857
—F1: 0.8831

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-05
—trainbatchsize: 16
—evalbatchsize: 128
—seed: 42
—gradientaccumulationsteps: 4
—totaltrainbatch_size: 64
—optimizer: Use adamwhf with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.5453	0.9933	112	0.3860	0.8277	0.8311	0.8314	0.8277
1.0577	1.9933	224	0.2918	0.8777	0.8843	0.8731	0.8757
0.6143	2.9933	336	0.3023	0.8876	0.8871	0.8888	0.8874
0.2438	3.9933	448	0.6792	0.8652	0.8714	0.8606	0.8630
0.063	4.9933	560	0.7500	0.8789	0.8817	0.8758	0.8776
0.052	5.9933	672	0.8892	0.8777	0.8832	0.8735	0.8758
0.0005	6.9933	784	0.9423	0.8801	0.8863	0.8758	0.8783
0.0002	7.9933	896	0.8404	0.8752	0.8777	0.8722	0.8738
0.0	8.9933	1008	0.8774	0.8777	0.8823	0.8738	0.8760
0.0	9.9933	1120	0.8828	0.8777	0.8823	0.8738	0.8760

Framework versions

—Transformers 4.48.2
—Pytorch 2.5.1+cu121
—Datasets 3.2.0
—Tokenizers 0.21.0