YOLO-NAS
YOLO-NAS is a computer vision model architecture developed by Deci AI.
Supported Model Types¶
You can deploy the following YOLO-NAS model types with Inference:
- Object Detection
- Keypoint Detection
Supported Inputs¶
Click a link below to see instructions on how to run a YOLO-NAS model on different inputs:
Configure Your Deployment¶
Starting from scratch? Use our Deployment Wizard to get a code snippet tailored to your device and use case.
- Configure a YOLO-NAS Object Detection Model
- Configure a YOLO-NAS Classification Model
- Configure a YOLO-NAS Segmentation Model
License¶
YOLO-NAS code is released under the Apache 2.0 license, but the Deci-provided pre-trained weights are under a special license.
- Training with Roboflow Train: Commercial usage is allowed because Roboflow does not use the Deci weights.
- Self-training outside Roboflow: If you train your own YOLO-NAS model outside the Roboflow platform, ensuring adherence to the Deci YOLO-NAS license is your responsibility.
See our Licensing Guide for more information about how your use of YOLO-NAS is licensed when using Inference to deploy your model.