baudm / parseq

Scene Text Recognition with Permuted Autoregressive Sequence Models (ECCV 2022)
https://huggingface.co/spaces/baudm/PARSeq-OCR
Apache License 2.0
565 stars 126 forks source link

After training #125

Closed RedSTwix closed 7 months ago

RedSTwix commented 9 months ago

I trained a new model, with a small dataset using pre-trained. Now I have the checkpoint and how can I load it into the project other than through read.py? and how can I reduce it, it is 363MB. There is very little information about this.