Hi, I'm Niels from the Hugging Face community team. I've opened this PR to add a descriptive model card for SAM3-LiteText.
This PR includes:
- Relevant metadata, including the `image-segmentation` pipeline tag.
- A summary of the model and its architecture (distilling SAM3's text encoder into MobileCLIP).
- Links to the research paper and the official GitHub repository.
- A sample usage snippet for text-prompted inference, taken directly from the project's documentation.
efficient sam3 image encoder mobileclip text encoders
