why??my result is not RGB

[pogevip]

[chunleiguo]

hi, maybe you need set the epoch a bigger number, say 1000, and try again. it works from my experience.

[developer0hye]

I also got the same results.

[developer0hye]

@pogevip In my case, this problem was solved when training the model with lower batch size like 4, 8, and 16.

[lordbutters]

@pogevip In my case, this problem was solved when training the model with lower batch size like 4, 8, and 16.

Sorry for disturb you, can you tell me the pytorch version, tensorflow version, cuda version and cudnn version that you have used when train the code? I got some problem in running the code

[developer0hye]

@lordbutters I used pytorch version 1.0 >= and 1.2 <=. maybe...

[CYTomGithub]

@pogevip In my case, this problem was solved when training the model with lower batch size like 4, 8, and 16.

sorry，In my case, I can not get the RGB picture when training the model with lower batch size like 4, 8, and 16. And I do not know why~

[Charrrrrlie]

In my case, the values of RGB channels in result picture come so close (e.g. 20/23/32) which contributes to its gray outlook.

[ZhangYuef]

In my case, the values of RGB channels in result picture come so close (e.g. 20/23/32) which contributes to its gray outlook.

Any advice to avoid this problem?

[Charrrrrlie]

In my case, the values of RGB channels in result picture come so close (e.g. 20/23/32) which contributes to its gray outlook.

Any advice to avoid this problem?

I think you can absorb the comments of chunleiguo and developerOhye in training FOR A BETTER WEIGHT. （I've done that months ago. I may not remember much detail.)

[ZhangYuef]

In my case, the values of RGB channels in result picture come so close (e.g. 20/23/32) which contributes to its gray outlook.

Any advice to avoid this problem?

I think you can absorb the comments of chunleiguo and developerOhye in training FOR A BETTER WEIGHT. （I've done that months ago. I may not remember much detail.)

Thanks for reply! I will try.