1/ In developing DeOldify, I've had a very helpful, yet seemingly obvious, rule of thumb to help guide figuring out improvements in training: What can be done to close the gap between training and inference? Getting critical about this can yield some seriously low hanging fruit.
-
Show this thread
2/ There's many, many things you can enumerate here. Here's a less obvious example: Are you account for jpeg artifacts? What about "moar jpeg" scenarios where you rejpeg an image to death? Turns out that's very common in the real world, of course, yet easy to overlook!
0 replies
1 retweet
25 likes
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.