Description I recently made a switch to 2.36.0 from 2.21.0 and noticed that some of my models have slowed down noticeably. I ran perf analyzer on the models in the 2 different versions and noticed that the newer version is able to atleas...
Since we are switch to cuda 12 as the default on official releases, we would need to swap all current packaging pipeline and publishing pipeline to set the cuda 12 as the default. Motivation and Context This is needed for current 1.19 release. -name:...
baijumeswanimerged commitf8af972intorel-0.3.0Jun 25, 2024 10 checks passed baijumeswanideleted thebaijumeswani/rel-0.3.0-cherrypickbranchJune 25, 2024 20:25 Sign up for freeto join this conversation on GitHub. Already have an account?Sign in to comment ...
Linux CPU x64 Build on: pull_request 2 Linux CPU ARM64 Build on: pull_request 1 Windows CPU arm64 Build on: pull_request 1 Linux CUDA x64 Build on: pull_request 2 Windows DirectML x64 Build on: pull_request Code scanning results Select a check to view from the sidebar Foot...