ModernBERT-large-roman-urdu-fine-grained

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: 1.4888
• Accuracy: 0.8088
• Precision: 0.7419
• Recall: 0.7023
• F1: 0.7205

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
• train_batch_size: 16
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use adamw_hf with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.3921	1.0	226	0.6527	0.7536	0.7638	0.5844	0.6408
1.0981	2.0	452	0.3654	0.8711	0.8685	0.7538	0.7984
0.7117	3.0	678	0.1802	0.9495	0.9303	0.9237	0.9266
0.1786	4.0	904	0.0550	0.9849	0.9789	0.9766	0.9777
0.1302	5.0	1130	0.0168	0.9950	0.9928	0.9927	0.9927
0.0194	6.0	1356	0.0062	0.9983	0.9970	0.9969	0.9970
0.0178	7.0	1582	0.0019	0.9997	0.9998	0.9995	0.9996
0.0003	8.0	1808	0.0004	0.9999	0.9999	0.9996	0.9997
0.0	9.0	2034	0.0006	0.9999	0.9999	0.9996	0.9997
0.0	9.9579	2250	0.0003	0.9999	0.9999	0.9996	0.9997

Framework versions

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