Hi, just starting out in this field, and amazing paper by the way!
I wanted to test this model on a custom dataset.
So I would like to know a bit about the rec field in your json annotations files. Could you please let me know how that kind of embedding is done in the literature?
The rec field sequentially records the indices of the characters in a word.
The word content can be recovered using the indices in the rec field and the character list in args.chars.
Hi, just starting out in this field, and amazing paper by the way! I wanted to test this model on a custom dataset. So I would like to know a bit about the rec field in your json annotations files. Could you please let me know how that kind of embedding is done in the literature?