gemini_chakma_banglaBert

This model is a fine-tuned version of sagorsarker/bangla-bert-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.9660

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
4.9302	1.0	161	4.4814
4.1569	2.0	322	4.1333
3.8372	3.0	483	4.0134
3.6273	4.0	644	3.8247
3.4985	5.0	805	3.6106
3.3996	6.0	966	3.6456
3.2831	7.0	1127	3.4770
3.1861	8.0	1288	3.4076
3.1303	9.0	1449	3.2995
3.0423	10.0	1610	3.2370
2.9555	11.0	1771	3.2081
2.9047	12.0	1932	3.1658
2.85	13.0	2093	3.0383
2.8092	14.0	2254	3.0543
2.7829	15.0	2415	3.0478
2.7282	16.0	2576	2.9925
2.7599	17.0	2737	2.9768
2.681	18.0	2898	2.9205
2.641	19.0	3059	2.9632
2.6854	20.0	3220	2.9660

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.0