Uploaded image for project: 'AI Platform Core Components'
  1. AI Platform Core Components
  2. AIPCC-5585

builder: flashinfer-cubin package update request

XMLWordPrintable

    • Icon: Epic Epic
    • Resolution: Duplicate
    • Icon: Normal Normal
    • None
    • None
    • Accelerator Enablement
    • flashinfer-cubin package onboarding
    • False
    • Hide

      None

      Show
      None
    • False
    • To Do
    • AIPCC-5592Build flashinfer-cubin wheels
    • 0% To Do, 0% In Progress, 100% Done

      Requested Package Name and Version:

      flashinfer-cubin

      Brief Explanation for request

      Right now, the flashinfer cubins are downloaded and installed manually:

      https://gitlab.com/redhat/rhel-ai/rhaiis/containers/-/blob/main/Containerfile.cuda-ubi9?ref_type=heads#L35-39

      It's not ideal and AIPCC should build the wheel instead, which would be shipped in the RHAIIS collection.

      QE user acceptance tests

      The wheel can be installed and cubins can correctly be used from VLLM.

      Package License

      Apache 2.0: https://pypi.org/project/flashinfer-cubin/ 

       

              emacchi@redhat.com Emilien Macchi
              emacchi@redhat.com Emilien Macchi
              Frank's Team
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated:
                Resolved: