I love this!
#13
by
Taylor-eOS
- opened
I couldn't get any OCR model to work. I had gone as far as writing a script that splits my input into single lines, and they were still getting it wrong.
Nanonets just works. Whole pages. Works. Other languages. Works.
If only it was MIT/Apache licensed :)
If only it was MIT/Apache licensed :)
We need the 7b model for that
Can you explain why MIT/Apache license matters?
Opens up the model to be used for basically anything -- the nanonets folks have claimed they are going to finetune the 7b model