CoolFace
Modelpublic

rikeshVertex/paddleocr-nepali-stage2-handwritten-fullft

sourceHugging Faceupdated 4mo agoView on Hugging Face
0likes3downloads
Model Card

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

paddleocr-nepali-stage2-handwritten-fullft

This model is a fine-tuned version of rikeshVertex/paddleocr-nepali-stage1-merged on an unknown dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.0998

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
  • —trainbatchsize: 4
  • —evalbatchsize: 4
  • —seed: 42
  • —gradientaccumulationsteps: 8
  • —totaltrainbatch_size: 32
  • —optimizer: Use OptimizerNames.PAGEDADAMW with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
  • —lrschedulertype: cosine
  • —lrschedulerwarmup_steps: 182
  • —training_steps: 1820

Training results

Training LossEpochStepValidation Loss
1.35940.21991000.8548
0.46570.43982000.5001
0.31910.65973000.3355
0.31350.87964000.2572
0.25431.09905000.2040
0.14281.31896000.1798
0.15961.53887000.1569
0.08501.75878000.1463
0.07411.97869000.1291
0.13492.197910000.1216
0.11492.417811000.1147
0.07872.637712000.1133
0.09542.857613000.1069
0.05983.077014000.1033
0.08093.296915000.1009
0.06133.516816000.1003
0.05963.736717000.0999
0.05813.956618000.0998
0.06164.018200.0998

Framework versions

  • —PEFT 0.19.1
  • —Transformers 5.7.0
  • —Pytorch 2.11.0+cu130
  • —Tokenizers 0.22.2