卡在cosine schedule这里不动了

[William20234]

No description provided.

[Algolzw]

你好，程序应该还是在跑，日志输出是单独保存在log文件夹里面的。

[melikaaala]

Hi I have the same issues and log file show ieltration but IT dosent move

[Algolzw]

Hi I have the same issues and log file show ieltration but IT dosent move

Hi you'd better run the code in the backend.

[melikaaala]

I have run in google colab. do you mean I should run on gpu?? thanks for your reply I really appreciate it

…

On Fri, May 3, 2024 at 4:51 PM balabala ***@***.***> wrote: Hi I have the same issues and log file show ieltration but IT dosent move Hi you'd better run the code in the backend. — Reply to this email directly, view it on GitHub <#75 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALIF3BGXCHNLUHNKCRPGEWTZAOFOLAVCNFSM6AAAAABDCU4MCSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAOJTGAYDSOJXHE> . You are receiving this because you commented.Message ID: ***@***.***>

[melikaaala]

You mean on GPU??

[Algolzw]

Yep, I mean the remote GPU server. If you want to train it on the colab, maybe you can explicitly print the loss by changing the logger.info(...) to print(...) here.

[melikaaala]

HI sorry to interrupt you but I have a problem with running your diffusion code. Does this log file mean it is running ?? but log stock in 400 and do not move.

…

On Sat, May 4, 2024 at 5:24 PM balabala ***@***.***> wrote: Yep, I mean the remote GPU server. If you want to train it on the colab, maybe you can explicitly print the loss by changing the logger.info(...) to print(...) here <https://github.com/Algolzw/image-restoration-sde/blob/2598d73cde3a1e2cf78c4ba4923414e1f4b90393/codes/config/deraining/train.py#L258> . — Reply to this email directly, view it on GitHub <#75 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALIF3BF5ID3G5P6MLJ4A7EDZATSDHAVCNFSM6AAAAABDCU4MCSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAOJUGIYDSNRVGU> . You are receiving this because you commented.Message ID: ***@***.***>

[Algolzw]

Hi! Have you solved this issue? You can check the GPU with nvidia-smi to see if the program is still running.