search

SamsungLabs / fbrs_interactive_segmentation

[CVPR2020] f-BRS: Rethinking Backpropagating Refinement for Interactive Segmentation https://arxiv.org/abs/2001.10331
Mozilla Public License 2.0
583 stars 94 forks source link

get oom error #37

Closed lfxx closed 3 years ago

lfxx commented 3 years ago

When i do inference in loop,gpu memory usage is growing continually which leads to oom error at last.How to resolve this problem?thank you!

ksofiyuk commented 3 years ago

It occurs due to the not very memory efficient GPU implementation of the DistMaps operator (memory consumption increases linearly with the number of clicks). However, we provide the memory and computational efficient implementation of this operator for the CPU. You can set cpu_dist_maps=True in the model to force using the CPU for this operator.