nepali_citiznship_model_final_with_belu

This model is a fine-tuned version of nielsr/lilt-xlm-roberta-base on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1526
  • Precision: 0.9148
  • Recall: 0.9151
  • F1: 0.9149
  • Accuracy: 0.9730
  • Bleu Score: 0.9619

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: 5e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • 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: 30

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy Bleu Score
No log 1.5625 50 0.3275 0.7670 0.8005 0.7834 0.9063 0.8494
No log 3.125 100 0.1563 0.8666 0.8883 0.8773 0.9602 0.9390
No log 4.6875 150 0.1267 0.8977 0.9098 0.9037 0.9695 0.9533
No log 6.25 200 0.1174 0.9105 0.9210 0.9157 0.9736 0.9618
No log 7.8125 250 0.1220 0.8984 0.9148 0.9065 0.9707 0.9591
No log 9.375 300 0.1240 0.9055 0.9151 0.9103 0.9720 0.9611
No log 10.9375 350 0.1175 0.9192 0.9174 0.9183 0.9742 0.9638
No log 12.5 400 0.1220 0.9050 0.9161 0.9105 0.9720 0.9611
No log 14.0625 450 0.1275 0.9202 0.9177 0.9190 0.9742 0.9635
0.1814 15.625 500 0.1307 0.9185 0.9164 0.9175 0.9740 0.9633
0.1814 17.1875 550 0.1347 0.9147 0.9214 0.9180 0.9741 0.9621
0.1814 18.75 600 0.1395 0.9168 0.9174 0.9171 0.9739 0.9629
0.1814 20.3125 650 0.1449 0.9224 0.9144 0.9184 0.9744 0.9631
0.1814 21.875 700 0.1473 0.9151 0.9118 0.9135 0.9729 0.9614
0.1814 23.4375 750 0.1448 0.9140 0.9164 0.9152 0.9731 0.9616
0.1814 25.0 800 0.1491 0.9109 0.9184 0.9146 0.9729 0.9619
0.1814 26.5625 850 0.1494 0.9121 0.9191 0.9156 0.9731 0.9622
0.1814 28.125 900 0.1539 0.9177 0.9101 0.9139 0.9732 0.9619
0.1814 29.6875 950 0.1526 0.9148 0.9151 0.9149 0.9730 0.9619

Framework versions

  • Transformers 4.50.3
  • Pytorch 2.6.0+cu124
  • Datasets 3.5.0
  • Tokenizers 0.21.1
Downloads last month
13
Safetensors
Model size
284M params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for bashyaldhiraj2067/nepali_citiznship_model_final_with_belu

Finetuned
(42)
this model