Logan Adams
Logan Adams
@janEbert - thanks, that does work for me. @crcrpar - could we get a review on and merge that PR? Also I don't need the changes to adding packaging and...
Hi @zhanwenchen - can you please update to the latest deepspeed and see if you still hit this issue? I wanted to see before trying to reproduce. Apologies for the...
Hi @zhanwenchen - closing this as stale for now, if you are still hitting this with the latest deepspeed, please comment and we can re-open and investigate.
@tohtana - curious if you have any thoughts on this?
@BSchilperoort - we can work on getting updated instructions in the Readme. The latest 0.17.5 release is broken for ROCm and we need to push a new update. I'll get...
> LGTM, but let's add a unit test to ensure this functionality can be tested on ROCm (and CUDA) @jagadish-amd - thoughts on adding unit tests for this?
> > > > LGTM, but let's add a unit test to ensure this functionality can be tested on ROCm (and CUDA) > > > > > > > >...
> @tjruwase Hi Olatunji, are there any further modifications that you'd like me to do? @traincheck-team - could you look into the failures on the nv-torch-latest-v100 test? We've stabilized the...
FYI @traincheck-team - there still appear to be some unit test errors if you could investigate.
FYI @traincheck-team - you'll need to sign off on the DCO requirement now that we've changed the GH organization/structure. Let me know if you have any questions on that.