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 of rejected SAM 3 access β lighthqg
Hello SAM 3 team,
My Hugging Face username is lighthqg. My application to access facebook/sam3 was rejected, and I would like to respectfully request a manual review.
I am conducting non-commercial academic research on vision-and-language navigation in simulated indoor environments using Habitat. Our project investigates how a navigation agent can more reliably identify instruction-relevant objects and verify its progress using visual and geometric evidence.
We would like to evaluate SAM 3 for text-prompted object detection, segmentation, and tracking. Specifically, we plan to compare its object grounding performance with that of our navigation VLM, and investigate whether its predictions can provide useful evidence for a separate navigation validator.
Our evaluation will examine missed detections, false positives, localization accuracy, temporal consistency, and inference latency. We will use appropriately licensed indoor-scene research datasets and comply with the applicable SAM 3 license terms and usage requirements.
Could you please reconsider my application? If any information needs clarification or correction, please let me know how I can provide it or update my request.
Thank you very much for your time and consideration.