NexaAI/yolo26n-npu

YOLO26n

Model Description

YOLO26n is the nano-sized variant of the YOLO26 model family, designed for ultra-fast, low-latency object detection on edge and resource-constrained devices.

As the smallest model in the YOLO26 lineup, YOLO26n prioritizes speed, simplicity, and deployability, while maintaining strong detection accuracy. It adopts an end-to-end, NMS-free architecture, which simplifies inference pipelines and reduces post-processing overhead.

YOLO26n is pretrained on the COCO dataset and serves as a lightweight baseline for real-time object detection workloads.

Quickstart

1. Install NexaSDK and create a free account at sdk.nexa.ai
2. Activate your device with your access token:

   nexa config set license '<access_token>'
   

3. Run the model on Qualcomm NPU in one line:

   nexa infer NexaAI/yolo26n-npu
   

Features

• Nano model size optimized for edge and low-power environments
• End-to-end detection with NMS-free inference
• Real-time capable for latency-sensitive applications
• Pretrained weights available out of the box
• Ultralytics workflow support for training, validation, inference, and export

Use Cases

• Edge and embedded computer vision
• Mobile object detection
• Smart cameras and IoT applications
• Robotics and autonomous systems
• Rapid prototyping with minimal compute resources

Inputs and Outputs

Input:

• Images or video streams, automatically preprocessed by the Ultralytics framework

Output:

• Bounding boxes
• Class labels
• Confidence scores

License

This repo is licensed under the Creative Commons Attribution–NonCommercial 4.0 (CC BY-NC 4.0) license, which allows use, sharing, and modification only for non-commercial purposes with proper attribution. All NPU-related models, runtimes, and code in this project are protected under this non-commercial license and cannot be used in any commercial or revenue-generating applications. Commercial licensing or enterprise usage requires a separate agreement. For inquiries, please contact dev@nexa.ai