You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
* Support GPU tensors
* Disable test_pipe.py to debug CI
* Add Request queues to Handle to support multiple asynchronous calls
* Enable distributed tests
* Remove debug print
rraminen
pushed a commit
to rraminen/DeepSpeed
that referenced
this issue
Apr 28, 2021
commit 7ebfd3fcce6008cf67ba85af4e8b7f5e24cea9b0 (HEAD, origin/xpu-main, origin/HEAD)
Author: Guo Yejun <yejun.guo@intel.com>
Date: Thu Oct 13 23:12:04 2022 +0800
pretrain_gpt2.py: do pytorch profiler at the last iteration (#23)
and don't add the last iteration time into average since the pt
profiler might impact performance.
DeepSpeed run should propagate python path just like we are doing NCCL vars.
The text was updated successfully, but these errors were encountered: