Why training needs 1000epoch?

yinboc / liif

Learning Continuous Image Representation with Local Implicit Image Function, in CVPR 2021 (Oral)

https://yinboc.github.io/liif/

BSD 3-Clause "New" or "Revised" License

1.28k stars 145 forks source link

Closed xiximelon closed 1 year ago

xiximelon commented 1 year ago

Actually I found at 500 epoch the training converges well, the results is similar to 1000 epoch