classify-document

Tool for classification of documents based on their layout.

Steps: gather some training data and label this by putting images in their respective folders. For example make two folders: letters and forms.

• ${traindir}/letters
• ${traindir}/forms

optionally split the training data into train and validation so you have the following folders:

• ${traindir}/letters
• ${traindir}/forms
• ${validationdir}/letters
• ${validationdir}/forms

Make sure none of the images are duplicated between train and validation

Train a model:

python3.6 main.py --do_train \ 
 --train_set ${traindir} \
 --do_validation \
 --validation_set ${validationdir} \
 --seed 42 \
 --gpu 0

Look at the numbers of the validation. They should be getting more correct each few epochs.

Inference a model:

The call will look like this:

python3 main.py --do_inference --inference_set /path/to/inference/dir/ --existing_model /path/to/model/best_val

if you have data that is not balanced (different numbers of items per class) it might make sense to add --use_class_weights