README.md

metadata

library_name: transformers
license: mit
base_model: ai4bharat/IndicBERTv2-MLM-Sam-TLM
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
model-index:
  - name: sutra-classifier-adv
    results: []

sutra-classifier-adv

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: 3.9104
• Accuracy: 0.9020
• F1: 0.9020
• Adv Acc: 0.0

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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Adv Acc
0.8968	1.0	72	1.9034	0.8078	0.8048	0.0902
2.4258	2.0	144	5.0244	0.8510	0.8503	0.1333
6.9155	3.0	216	8.0457	0.8549	0.8542	0.0
7.4505	4.0	288	12.3074	0.8706	0.8706	0.0
10.0671	5.0	360	9.0296	0.8353	0.8334	0.0
8.9089	6.0	432	10.0976	0.8275	0.8274	0.0353
6.8576	7.0	504	6.8378	0.8824	0.8824	0.0118
5.9407	8.0	576	5.4805	0.8588	0.8583	0.0314
5.1865	9.0	648	4.0087	0.8745	0.8745	0.1412
3.5606	10.0	720	4.4264	0.8549	0.8546	0.0078
3.4944	11.0	792	3.9312	0.9098	0.9098	0.0
3.4093	12.0	864	4.0868	0.8902	0.8902	0.0
3.3966	13.0	936	3.9650	0.8980	0.8979	0.0
3.3191	14.0	1008	4.0058	0.8824	0.8821	0.0
3.3156	15.0	1080	4.0279	0.8784	0.8784	0.0
3.26	16.0	1152	4.0243	0.8902	0.8902	0.0
3.264	17.0	1224	3.9404	0.8941	0.8941	0.0
3.2533	18.0	1296	3.8867	0.8941	0.8941	0.0
3.2137	19.0	1368	3.8939	0.9020	0.9020	0.0
3.2148	20.0	1440	3.9104	0.9020	0.9020	0.0

Framework versions

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