Bump up the CUDA OOM test memory size (#48029)
Summary:
80GB is no longer large any more https://nvidianews.nvidia.com/news/nvidia-doubles-down-announces-a100-80gb-gpu-supercharging-worlds-most-powerful-gpu-for-ai-supercomputing
Hopefully, the new size could be OK until the end of Moore's Law :)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48029
Reviewed By: linbinyu
Differential Revision: D25003603
Pulled By: zou3519
fbshipit-source-id: 626b9c031daee950df8453be4d7643dd67647213