Conversation
* Fix AmpCast for float16 OneDNN doesn't support float16 format so fallback to standard implementation is needed. It fixes issue 19631. * Enable amp_cast test for float16 on CPU context
|
Hey @anko-intel , Thanks for submitting the PR
CI supported jobs: [sanity, windows-gpu, unix-cpu, unix-gpu, website, edge, clang, centos-cpu, centos-gpu, windows-cpu, miscellaneous] Note: |
|
@mxnet-bot run ci [centos-gpu, unix-cpu, windows-cpu] |
|
Jenkins CI successfully triggered : [centos-gpu, unix-cpu, windows-cpu] |
|
@mxnet-bot run ci [macosx-x86_64, centos-gpu, windows-cpu] |
|
Jenkins CI successfully triggered : [centos-gpu, windows-cpu] |
|
@mxnet-bot run ci [windows-gpu, windows-cpu] |
|
Jenkins CI successfully triggered : [windows-cpu, windows-gpu] |
|
@szha, could you review and merge this forward port from v1.x fix |
Forward port #19749 from v1.x to master.
OneDNN doesn't support float16 format so fallback to standard
implementation is needed.
It fixes issue #19631.