- 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
- Codes for flask api
- Infer VQA model
- Codes for chrome extension api
- Not yet done with inferring model due to problem with onnx. Details will be in report.
- we refered to this git: SAM_TEXTVQA
- you can download best_model.tar about our experiments from this link
- paper: link
- Code to convert pytorch checkpoint to onnx
- Works well converting one model. Details with onnx will be in report.
- Code to implement TextVQA demo
- Need Google paid API to make OCR tokens
- Codes to connect Javascript file and python file using ajax and flask