jytime / Mask_RCNN_Pytorch

Mask R-CNN for object detection and instance segmentation on Pytorch

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Run demo_coco.py appear Cuda out of memory

InstantWindy opened this issue · comments

Hi! Does it take up a lot of memory to test demo_coco files?I don't know that every time I run this file there is always a Cuda out of memory. Thanks!

hi
it will occupy around 10 GB. Please notice the settings if it reports 'out of memory'

Can I ask you a question?When my convolution kernel value is only 1 and -1 or when my input value is only 1 and -1, will the network run faster?

It's a bit weird. Do you mean change the kernel size? If so, you have to re-design the network architecture and cannot use the pretrained model.