I have tried running this on the MVTec dataset whilst using the wide-resnet-50 and layer2 and layer3. I am running this on a machine with 64gb of RAM (fairly big). However I get the error:
"DefaultCPUAllocator: not enough memory: you tried to allocate 107395154944 bytes"
This is a huge RAM requirement. Am I doing anything wrong - how have other people got around this?
This was happening because I was running the sampler "greedy_coreset" instead of "approx_greedy_coreset". Leaving this here incase it helps anyone else.
I have tried running this on the MVTec dataset whilst using the wide-resnet-50 and layer2 and layer3. I am running this on a machine with 64gb of RAM (fairly big). However I get the error: "DefaultCPUAllocator: not enough memory: you tried to allocate 107395154944 bytes" This is a huge RAM requirement. Am I doing anything wrong - how have other people got around this?
Thanks