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")# pip install -U transformers accelerate # 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
Hello SAM 3 Team,
I am requesting access to facebook/sam3 for non-commercial academic research and reproduction.
Project: OVOW (One Video, One World), an open-source video-to-3D/world reconstruction project.
Use case:
SAM 3 is used in the official OVOW pipeline for per-frame object segmentation, providing object masks for subsequent reconstruction and evaluation stages. I am reproducing the official pipeline for academic research and experimental evaluation.
The model weights will be downloaded only through the official Hugging Face Hub using an authenticated token and used locally for research inference. I will comply with Meta’s model license and gated-access terms, and I will not redistribute the weights, host public mirrors, or use the model for prohibited purposes.
Affiliation: Shanghai Jiao Tong University
Hugging Face username: Xintong-Dong
Contact email: jesidongxt@sjtu.edu.cn
My previous access request was rejected. Could you please manually review or reset my access request so that I can reproduce the officially documented OVOW workflow?
Thank you for your time and consideration.