Form Classifier

Classification model for historical handwritten documents.

TO-DO

• Find a model pre-trained on a dataset closer to our domain
• Include embeddings of words extracted from first iteration of ocr tool (kraken)

Correct # In[0] comments with:

sed -i.bak -e "s/\#\s*In\[[0-9 ]*\]/#In\[\]/" Prototype.py