Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
7moritz7
on May 29, 2023
|
parent
|
context
|
favorite
| on:
Donut: OCR-Free Document Understanding Transformer
The non quantized models look relatively large, 800 MB. So you'd probably need to do inference on a server and somehow monetize that. Sounds difficult
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: