bytedance / sa2va-26b-image

Sa2VA: Marrying SAM2 with LLaVA for Dense Grounded Understanding of Images and Videos

  • Public
  • 1.1K runs
  • GitHub
  • Weights
  • Paper
  • License
Run with an API
Playground API Examples README Versions
  1. Author
    @8w9ag
    Version
    sha-67ecb2d
    Commit
    c859c181a6adba25e84547497b5842a84adde46e

    addd35cc

    Latest

    Release notes

    fast-push

  2. Author
    @lucataco
    Version
    cuda12.4-python3.10-torch2.4.0-X64
    Commit
    6f59e55da541fad1ffc370419265eab6f511569f

    9ed0fc32

    Release notes

    fixed masking

  3. Author
    @lucataco
    Version
    cuda12.4-python3.10-torch2.4.0-X64
    Commit
    6f59e55da541fad1ffc370419265eab6f511569f

    34bd6800