Edit model card

layoutlmv3-finetuned-invoice

This model is a fine-tuned version of microsoft/layoutlmv3-base on the sroie dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0030
  • Precision: 1.0
  • Recall: 0.9980
  • F1: 0.9990
  • Accuracy: 0.9998

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: 1e-05
  • train_batch_size: 2
  • eval_batch_size: 2
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • training_steps: 2000

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
No log 2.0 100 0.0715 0.972 0.9858 0.9789 0.9971
No log 4.0 200 0.0228 0.972 0.9858 0.9789 0.9971
No log 6.0 300 0.0174 0.972 0.9858 0.9789 0.9971
No log 8.0 400 0.0137 0.972 0.9858 0.9789 0.9971
0.1189 10.0 500 0.0122 0.972 0.9858 0.9789 0.9971
0.1189 12.0 600 0.0112 0.972 0.9858 0.9789 0.9971
0.1189 14.0 700 0.0080 0.972 0.9858 0.9789 0.9971
0.1189 16.0 800 0.0100 0.972 0.9858 0.9789 0.9971
0.1189 18.0 900 0.0040 0.9960 0.9980 0.9970 0.9996
0.0097 20.0 1000 0.0030 1.0 0.9980 0.9990 0.9998
0.0097 22.0 1100 0.0028 0.9980 0.9959 0.9970 0.9996
0.0097 24.0 1200 0.0016 1.0 1.0 1.0 1.0
0.0097 26.0 1300 0.0015 1.0 1.0 1.0 1.0
0.0097 28.0 1400 0.0015 0.9980 0.9980 0.9980 0.9998
0.0029 30.0 1500 0.0017 0.9980 0.9980 0.9980 0.9998
0.0029 32.0 1600 0.0026 0.9960 0.9980 0.9970 0.9996
0.0029 34.0 1700 0.0026 0.9960 0.9980 0.9970 0.9996
0.0029 36.0 1800 0.0026 0.9960 0.9980 0.9970 0.9996
0.0029 38.0 1900 0.0025 0.9960 0.9980 0.9970 0.9996
0.002 40.0 2000 0.0026 0.9960 0.9980 0.9970 0.9996

Framework versions

  • Transformers 4.20.0.dev0
  • Pytorch 1.11.0+cu113
  • Datasets 2.2.2
  • Tokenizers 0.12.1
Downloads last month
14
Hosted inference API
Token Classification
Examples
Examples
This model can be loaded on the Inference API on-demand.

Evaluation results