The code repo for ICDAR 2023 paper: Open-Set Text Recognition via Shape-Awareness Visual Reconstruction
[UPD] The models are now uploaded to google drive, at least it does not ask my phone number for uploading and won't ask yours for downloading...
Link: https://drive.google.com/file/d/1EhaAbKEWtKUUivEmb_0mbrWDKvimvepz/view?usp=sharing
md5sum: 9bc5335daf740e8129eb6b0bdc82b216 opensavr-models.zip
The cat is still finding a new sharing site where everyone can download without having to pay, register, or install nonsense shenanigans. If you have a better hosting choice, please share it with us.