ConsistentTeacher icon indicating copy to clipboard operation
ConsistentTeacher copied to clipboard
verified publisher icon Adamdad

8 * v100 parallel training

Open Re-dot-art opened this issue 1 year ago • 3 comments

I can train normally on 2v100, but the following error occurred on 8v100. Have you encountered it? Thanks!

image

Re-dot-art avatar Mar 19 '24 06:03 Re-dot-art

Sorry, but this seems not to be a problem of our code but a problem of multiprocessing in python. Could you please provide details about your running environment, the commands you're using, and any shell scripts involved?

Adamdad avatar Mar 19 '24 06:03 Adamdad

Sorry, but this seems not to be a problem of our code but a problem of multiprocessing in python. Could you please provide details about your running environment, the commands you're using, and any shell scripts involved?

image cuda=12.0,python=3.8, I ran the coco fully label experimental setup on 8 GPUs.

Re-dot-art avatar Mar 20 '24 09:03 Re-dot-art

Sorry, but this seems not to be a problem of our code but a problem of multiprocessing in python. Could you please provide details about your running environment, the commands you're using, and any shell scripts involved?

image

Re-dot-art avatar Mar 20 '24 09:03 Re-dot-art