Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

What a terrific project! So many levels of challenges. How many different images of each piece was needed to before you could train the system?


Some I now have 100's of images of some only a few tens or even just a few. Obviously the error rates for the latter are higher.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: