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 Manual Review / Reset of Rejected SAM 3 Access — RoboHarness Reproduction Research
Hello Meta SAM 3 Team,
My access request to facebook/sam3 was rejected, and I would like to kindly request a manual review or a reset of my access status.
I am an embodied AI researcher currently working on reproducing the paper “RoboHarness: A Memory-Augmented Policy Harness for Vision-Language-Action Model Robustness via In-Context Adaptation” (arXiv:2603.24060, IROS 2026).
In the official RoboHarness implementation, SAM 3 is used as part of the visual perception / MCP tool pipeline. The released code includes a sam3_service.py service and SAM3-related visual tool calls used by the RoboHarness agent during robot-policy evaluation.
I would like to use the official SAM 3 model weights to faithfully reproduce the published RoboHarness pipeline, evaluate its visual intervention mechanism, and conduct research experiments on embodied AI and Vision-Language-Action systems.
The model will be used locally in my research environment. I will comply with the SAM 3 license, Meta policies, and Hugging Face terms, and I will not redistribute the model weights or my access credentials.
Could you please manually review my rejected request, or reset my access status so that I can submit the gated-access application again?
Hugging Face username: robinson69
Affiliation: [NorthEastern University]
Role: Embodied AI Researcher
Research purpose: Reproduction and evaluation of RoboHarness and its SAM 3-based visual perception/MCP pipeline
I would be happy to provide any additional information if needed.
Thank you very much for your time and consideration.
Best regards,
[robinson]