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
We are currently working on support for HPU hardware on the training side of DeepSpeed. I expect we will also work towards supporting it for inference as well, but we do not have any set goals/deadlines around that support at this time.
Currently the TGI have tgi-gaudi to support the HPU for Gaudi 2 and also VLLM is included their roadmap in Q1 2024
Is the DeepSpeed-MII will support habana (HPU) hardware as well?
Thank you
Tasks
The text was updated successfully, but these errors were encountered: