sutra-classifier-adv-epoch50

This model is a fine-tuned version of ai4bharat/IndicBERTv2-MLM-Sam-TLM on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.9950
• Accuracy: 0.8941
• F1: 0.8938
• Adv Acc: 0.0196

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: 8.8e-05
• train_batch_size: 32
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.079905
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Adv Acc
0.783	1.0	72	0.7087	0.8549	0.8549	0.1451
0.7764	2.0	144	3.0179	0.8863	0.8862	0.1451
2.9074	3.0	216	2.7534	0.8941	0.8941	0.1451
3.3278	4.0	288	4.3535	0.8549	0.8546	0.0
3.8347	5.0	360	5.6863	0.8902	0.8902	0.1216
5.5114	6.0	432	5.2154	0.5098	0.3926	0.1098
4.1185	7.0	504	8.2343	0.8745	0.8745	0.0706
6.9436	8.0	576	3.4735	0.8549	0.8547	0.0275
3.3697	9.0	648	3.7890	0.8745	0.8745	0.1059
3.151	10.0	720	3.8379	0.8667	0.8664	0.1961
2.8458	11.0	792	3.4405	0.9137	0.9136	0.0039
3.0388	12.0	864	3.4150	0.8941	0.8941	0.0039
3.0298	13.0	936	3.8170	0.9059	0.9054	0.0
3.1019	14.0	1008	4.0375	0.8745	0.8745	0.0
3.1136	15.0	1080	4.3493	0.8627	0.8624	0.0510
3.1902	16.0	1152	4.1109	0.8745	0.8744	0.0078
3.2377	17.0	1224	4.1627	0.8588	0.8586	0.0
3.2451	18.0	1296	4.3464	0.8706	0.8694	0.0
3.2331	19.0	1368	4.0302	0.8471	0.8469	0.0235
3.2224	20.0	1440	4.2660	0.8784	0.8784	0.0
3.3011	21.0	1512	3.9950	0.8941	0.8938	0.0196

Framework versions

• Transformers 4.49.0
• Pytorch 2.7.1+cu126
• Datasets 4.0.0
• Tokenizers 0.21.4

WandB training stats

https://wandb.ai/bhaswata08/huggingface/runs/p5ma54yk/workspace