-
Notifications
You must be signed in to change notification settings - Fork 7.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
ColTran OOM Error even with small dataset #838
Comments
I hardcoded the battch-size for the spatial upsampler here to be 5. Could you change it and see if it works? (https://github.com/google-research/google-research/blob/master/coltran/custom_colorize.py#L168) You could also set the |
Did this fix your issue? Thanks! |
Yes, I missed updating it here. |
hi i am having the same probelm with spatial upsampler part, it gives OOM so i have to run it on cpu by uncommenting the line 6 of custom_colorise, i get following error with gpu
i did try batch size 1 in custom_colorize still same error |
@haiderasad Can you try the hack mentioned in this blogpost? (https://habr.com/ru/company/ruvds/blog/563858/)
You can save some memory by changing from float32 to float16. |
I am getting the OOM error (Out Of Memory) on running the script for Coltran spatial upsampler.
Getting the same error for the imagenet dataset of 100, 100, 50 and even 10 images for both modes (colorize, recolorize).
The
/configs/spatial_upsampler.py
file already hasconfig.batch_size = 1
, so cannot reduce further.The first 2 steps colourizer and color_upsampler are running fine.
Command used
env variables
I am running this on Google Colab
nvidia-smi
Error
Please guide me on how to solve this.
The text was updated successfully, but these errors were encountered: