bytedance/sa2va-26b-image

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

Public
5.6K runs
  1. Author
    @8w9ag
    Version
    sha-67ecb2d

    addd35cc

    Latest

    Release notes

    fast-push

  2. Release notes

    fixed masking