Large scene with 8000 images
Hello everyone, I have a video shot in a large indoor scene, which has generated over 8,000 images, extracted in sequence from the video. I'm wondering if current systems support processing such a large number of images? It seems like the GPU memory might not be sufficient.
Maybe mini-batch optimization + gradient accumulation will help? currently the code using full-batch so is not scalable.
Oh, okay. Maybe this is a direction worth trying. Have you also been researching this project recently?
Maybe mini-batch optimization + gradient accumulation will help? currently the code using full-batch so is not scalable.
no, but it is easy to try, let me know if it works, you can try this on official small dataset first to see if accuracy degrade.