CS470

• CS470 final project for Team 32
• Implementation of VQA and TextVQA for blind people
• Details of each folder will be inside README of each folder

API folder

• Codes for flask api
• Infer VQA model

Web folder

• Codes for chrome extension api
• Not yet done with inferring model due to problem with onnx. Details will be in report.

TextVQA folder

• we refered to this git: SAM_TEXTVQA
• you can download best_model.tar about our experiments from this link
• paper: link

convert_pt2onnx.py

• Code to convert pytorch checkpoint to onnx
• Works well converting one model. Details with onnx will be in report.

RunModel folder(failed)

• Code to implement TextVQA demo
• Need Google paid API to make OCR tokens

ajax-flask(failed)

• Codes to connect Javascript file and python file using ajax and flask