pytorch

86fbf17b - Use std::vector instead of alloca to work around hcc crash

Use std::vector instead of alloca to work around hcc crash Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/15175 Differential Revision: D13453708 Pulled By: bddppq fbshipit-source-id: f8c147ae9f679e395fee9d4c73ebcca052c9a752