YOLO26-Detection: Optimized for Qualcomm Devices
Ultralytics YOLO26 is a machine learning model that predicts bounding boxes and classes of objects in an image.
This is based on the implementation of YOLO26-Detection found here. This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the Qualcomm® AI Hub Models library to export with custom configurations. More details on model performance across various devices, can be found here.
Qualcomm AI Hub Models uses Qualcomm AI Hub Workbench to compile, profile, and evaluate this model. Sign up to run these models on a hosted Qualcomm® device.
Getting Started
Due to licensing restrictions, we cannot distribute pre-exported model assets for this model. Use the Qualcomm® AI Hub Models Python library to compile and export the model with your own:
- Custom weights (e.g., fine-tuned checkpoints)
- Custom input shapes
- Target device and runtime configurations
See our repository for YOLO26-Detection on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.object_detection
Model Stats:
- Model checkpoint: YOLO26-N
- Input resolution: 640x640
- Number of parameters: 2.4M
- Model size (float): 9.2 MB
- Model size (w8a16): 3.2 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| YOLO26-Detection | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.344 ms | 0 - 81 MB | NPU |
| YOLO26-Detection | ONNX | w8a16 | Snapdragon® X2 Elite | 2.512 ms | 0 - 0 MB | NPU |
| YOLO26-Detection | ONNX | w8a16 | Snapdragon® X Elite | 6.143 ms | 2 - 2 MB | NPU |
| YOLO26-Detection | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.466 ms | 0 - 225 MB | NPU |
| YOLO26-Detection | ONNX | w8a16 | Qualcomm® QCS6490 | 323.555 ms | 99 - 105 MB | CPU |
| YOLO26-Detection | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 5.589 ms | 2 - 7 MB | NPU |
| YOLO26-Detection | ONNX | w8a16 | Qualcomm® QCS9075 | 6.316 ms | 2 - 5 MB | NPU |
| YOLO26-Detection | ONNX | w8a16 | Qualcomm® QCM6690 | 154.87 ms | 101 - 111 MB | CPU |
| YOLO26-Detection | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.691 ms | 0 - 75 MB | NPU |
| YOLO26-Detection | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 135.547 ms | 103 - 112 MB | CPU |
| YOLO26-Detection | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.144 ms | 1 - 162 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Snapdragon® X2 Elite | 2.816 ms | 5 - 5 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Snapdragon® X Elite | 4.736 ms | 5 - 5 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 3.167 ms | 5 - 181 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 12.913 ms | 1 - 156 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 4.328 ms | 5 - 6 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Qualcomm® SA8775P | 5.731 ms | 0 - 157 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Qualcomm® QCS9075 | 6.153 ms | 7 - 13 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 8.682 ms | 5 - 197 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Qualcomm® SA7255P | 12.913 ms | 1 - 156 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Qualcomm® SA8295P | 9.232 ms | 0 - 168 MB | NPU |
| YOLO26-Detection | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.431 ms | 5 - 165 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.004 ms | 1 - 183 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 2.375 ms | 2 - 2 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Snapdragon® X Elite | 4.93 ms | 2 - 2 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.115 ms | 2 - 205 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Qualcomm® QCS8275 (Proxy) | 8.251 ms | 1 - 177 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 4.557 ms | 2 - 4 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Qualcomm® SA8775P | 5.24 ms | 1 - 182 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 5.143 ms | 1 - 5 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Qualcomm® QCM6690 | 19.586 ms | 2 - 180 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Qualcomm® SA7255P | 8.251 ms | 1 - 177 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.348 ms | 2 - 182 MB | NPU |
| YOLO26-Detection | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 5.084 ms | 2 - 183 MB | NPU |
License
- The license for the original implementation of YOLO26-Detection can be found here.
References
- Ultralytics YOLO26: NMS-Free Real-Time Object Detection for Edge Devices
- Source Model Implementation
Community
- Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
- For questions or feedback please reach out to us.