mahabert_output

This model is a fine-tuned version of l3cube-pune/marathi-bert-v2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6570
• Accuracy: 0.815
• F1: 0.7933
• Precision: 0.8452
• Recall: 0.7474

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• 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
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
No log	1.0	25	0.6905	0.625	0.6914	0.5676	0.8842
No log	2.0	50	0.6701	0.795	0.7684	0.8293	0.7158
No log	3.0	75	0.6570	0.815	0.7933	0.8452	0.7474

Framework versions

• Transformers 4.53.3
• Pytorch 2.6.0+cu124
• Datasets 4.4.1
• Tokenizers 0.21.2