Logan Adams

Results 466 comments of Logan Adams

[BUG] Pipeline Dataloader Sampler: `shuffle=False`

Thanks @xianshunw and @Coobiw - we will work on making it configurable, but at least with the current unit tests, the linked PR seems to hang with `shuffle=True`, so we...

[BUG] Pipeline Dataloader Sampler: `shuffle=False`

Creating #6950 to track adding the config option.

How to silence warning when import deepseed without change the source code?

@sean-wade - this PR has done some work on this, would this be enough for now? https://github.com/microsoft/DeepSpeed/pull/5917

How to silence warning when import deepseed without change the source code?

Closing this for now as the PR above is merged

[BUG] File not found in autotuner cache in multi-node setting on SLURM

@jubueche - the related PR is now resolved, can you see if you are still hitting this if you use the latest DeepSpeed?

[BUG] File not found in autotuner cache in multi-node setting on SLURM

@ClaartjeBarkhofTNO and @jubueche - thanks for the update on this, we will take a look. Can you share your DeepSpeed version as well as any other info about your system...

when I finetune the model use deepspeed on 2 4*A800s，log only contain worker1

@bill4689 - following up on this if you have any updates?

[BUG] oneapi/ccl.hpp: No such file or directory.

Thanks @weiji14 for opening this to track.

[BUG] oneapi/ccl.hpp: No such file or directory.

@weiji14 - this should be fine to add to the dependencies, it should not cause any issues on the CUDA builds. Also it should be fine to leave `DS_BUILD_OPS=1`, That...

[BUG] oneapi/ccl.hpp: No such file or directory.

@whois206 - I had no problems running the command that you listed: ``` (base) test@deepspeed:~$ conda install -c https://software.repos.intel.com/python/conda/ -c conda-forge oneccl-devel Channels: - https://software.repos.intel.com/python/conda - conda-forge - defaults Platform:...