Document_IE
Document_IE copied to clipboard
Questions.
Hi! First of all, thanks for this repo. It is very insightful!
Do you have the code you preprocessed SROIE data for using this model? Can you provide a guide to reproduce your results?
Hi! First of all, thanks for this repo. It is very insightful!
Do you have the code you preprocessed SROIE data for using this model? Can you provide a guide to reproduce your results?
I'm really sorry for replying late. My work has switched to LLM,and busy with it now. In my mind, The origin SROIE data comes from a competetion with OCR box information, such as box position and text content. First, use box information as input of grapher.py to generate graph matrixs. Second, use graph matrixs as input of train_lstm_gcn.py to train model.