tes_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: 1.8596

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
2.9436	1.0	465	2.4593
2.0494	2.0	930	2.2688
1.8839	3.0	1395	2.2129
1.8058	4.0	1860	2.1472
1.7363	5.0	2325	2.0998
1.6849	6.0	2790	2.0739
1.64	7.0	3255	2.0289
1.5953	8.0	3720	1.9871
1.5632	9.0	4185	1.9855
1.5379	10.0	4650	1.9571
1.5179	11.0	5115	1.9333
1.4938	12.0	5580	1.9415
1.4686	13.0	6045	1.8951
1.4604	14.0	6510	1.8786
1.4409	15.0	6975	1.8658
1.4287	16.0	7440	1.8866
1.4099	17.0	7905	1.8704
1.4014	18.0	8370	1.8684
1.3986	19.0	8835	1.8314
1.3956	20.0	9300	1.8596

Framework versions

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