gemini_chakma_deberta

This model is a fine-tuned version of microsoft/deberta-v3-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.7292

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
3.0723	1.0	406	2.5866
2.0371	2.0	812	2.3154
1.8282	3.0	1218	2.1631
1.6799	4.0	1624	2.0851
1.5833	5.0	2030	1.9973
1.5115	6.0	2436	1.9709
1.4533	7.0	2842	1.9600
1.4026	8.0	3248	1.8901
1.3613	9.0	3654	1.8607
1.3392	10.0	4060	1.8564
1.3192	11.0	4466	1.8507
1.2828	12.0	4872	1.8288
1.2612	13.0	5278	1.7719
1.2493	14.0	5684	1.7531
1.2302	15.0	6090	1.7606
1.2195	16.0	6496	1.7127
1.2102	17.0	6902	1.7352
1.1999	18.0	7308	1.6995
1.1824	19.0	7714	1.7122
1.1919	20.0	8120	1.7292

Framework versions

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