librarian-bot's picture
Update README.md
daad568 verified
|
raw
history blame
2.32 kB
metadata
library_name: transformers
license: apache-2.0
base_model:
  - lightonai/LightOnOCR-1B-1025
tags:
  - generated_from_trainer
model-index:
  - name: lightonocr-books-test
    results: []

lightonocr-books-test

This model is a fine-tuned version of lightonai/LightOnOCR-1B-1025 on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1898

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: 6e-05
  • train_batch_size: 4
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 16
  • 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
  • lr_scheduler_warmup_steps: 10
  • num_epochs: 1

Training results

Training Loss Epoch Step Validation Loss
0.9005 0.0619 50 0.2190
0.7568 0.1239 100 0.2133
0.7556 0.1858 150 0.2129
0.8055 0.2478 200 0.2108
0.7609 0.3097 250 0.2021
0.6656 0.3716 300 0.1993
0.7036 0.4336 350 0.1979
0.8552 0.4955 400 0.1929
0.8087 0.5574 450 0.1944
0.7815 0.6194 500 0.1913
0.6349 0.6813 550 0.1930
0.7542 0.7433 600 0.1901
0.6332 0.8052 650 0.1913
0.6949 0.8671 700 0.1904
0.7974 0.9291 750 0.1896
0.5963 0.9910 800 0.1898

Framework versions

  • Transformers 5.0.0.dev0
  • Pytorch 2.9.0+cu128
  • Datasets 4.4.0
  • Tokenizers 0.22.1