Uploaded image for project: 'AI Platform Core Components'
  1. AI Platform Core Components
  2. AIPCC-680

Build vllm With Support for Audio and Video Models

XMLWordPrintable

    • False
    • Hide

      There are licensing concerns with several of the packages needed for this ticket.

      • AIPCC-816
      • AIPCC-818
      • AIPCC-820

       

      Show
      There are licensing concerns with several of the packages needed for this ticket. AIPCC-816 AIPCC-818 AIPCC-820  
    • True
    • 0% To Do, 0% In Progress, 100% Done
    • Hide

      17/Sept/25 - Green

      AIPCC handing off to RHAIIS team for validation.

      Show
      17/Sept/25 - Green AIPCC handing off to RHAIIS team for validation.

      Via RHOAIRFE-532, RHOAI wants to support audio and video models. We need to build the dependencies.

      From the RFE ticket:

      A vLLM image should enable deployment and usage of video/audio models. Below are python packages required by the above models to process video and audio, currently not build into the vllm images.

      librosa                           0.10.2.post1
soundfile                         0.13.0
soxr                              0.5.0.post1

       

           "audio": ["librosa", "soxr", "soundfile"],  # Required for audio processing

              mdean@redhat.com Meirav Dean
              dhellman@redhat.com Doug Hellmann
              Votes:
              0 Vote for this issue
              Watchers:
              13 Start watching this issue

                Created:
                Updated:
                Resolved: