Instructions to use facebook/sam3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use facebook/sam3 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("mask-generation", model="facebook/sam3")# Load model directly from transformers import AutoProcessor, AutoModel processor = AutoProcessor.from_pretrained("facebook/sam3") model = AutoModel.from_pretrained("facebook/sam3", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Request for reconsideration of SAM 3 gated model access
Hello SAM 3 team,
My access request to the SAM 3 gated repository was recently rejected. I believe my initial application may have been incomplete or did not clearly explain my intended use.
I am evaluating SAM 3 for an academic and engineering research project focused on interactive image segmentation for stylized 2D characters.
My planned evaluation includes:
- text-prompt segmentation
- point-prompt segmentation
- box-prompt segmentation
- semantic part segmentation such as face, eyes, arms, body, and accessories
- robustness testing across different character styles, poses, and backgrounds
- inference latency and GPU memory measurements
The experiments will be performed locally on an NVIDIA RTX 3090 workstation. The goal is to evaluate whether SAM 3 can support an interactive character-authoring workflow where users can select and refine visual parts before creating animation assets.
This is a research and prototyping evaluation only. I will comply with the SAM 3 license and Hugging Face terms. I will not redistribute the model weights, expose them as a public model service, or attempt to bypass any access restrictions.
Could you please reset my rejected access request so that I can submit the application again with complete information, or reconsider the existing request?
Thank you for your time and help.