YOLOv11-Pose: Optimized for Qualcomm Devices
Ultralytics YOLOv11-Pose is a machine learning model that predicts human body keypoints and bounding boxes in an image.
This is based on the implementation of YOLOv11-Pose 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 YOLOv11-Pose on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.pose_estimation
Model Stats:
- Input resolution: 640x640
- Model checkpoint: YOLO11-N-Pose
- Model size (float): 11.1 MB
- Model size (w8a16): 3.63 MB
- Number of parameters: 2.9M
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| YOLOv11-Pose | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.958 ms | 3 - 222 MB | NPU |
| YOLOv11-Pose | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.348 ms | 0 - 206 MB | NPU |
| YOLOv11-Pose | ONNX | float | Snapdragon® X2 Elite | 3.249 ms | 9 - 9 MB | NPU |
| YOLOv11-Pose | ONNX | float | Snapdragon® X Elite | 6.517 ms | 9 - 9 MB | NPU |
| YOLOv11-Pose | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 4.019 ms | 8 - 230 MB | NPU |
| YOLOv11-Pose | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 11.221 ms | 8 - 215 MB | NPU |
| YOLOv11-Pose | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 6.901 ms | 6 - 10 MB | NPU |
| YOLOv11-Pose | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 6.154 ms | 6 - 12 MB | NPU |
| YOLOv11-Pose | ONNX | float | Qualcomm® QCS8450 | 11.221 ms | 8 - 215 MB | NPU |
| YOLOv11-Pose | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 7.567 ms | 6 - 10 MB | NPU |
| YOLOv11-Pose | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 6.517 ms | 9 - 9 MB | NPU |
| YOLOv11-Pose | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 3.348 ms | 0 - 206 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.419 ms | 0 - 93 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.68 ms | 1 - 89 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Snapdragon® X2 Elite | 2.549 ms | 4 - 4 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Snapdragon® X Elite | 6.105 ms | 3 - 3 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.674 ms | 4 - 226 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 7.206 ms | 4 - 223 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 21.068 ms | 3 - 7 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 5.301 ms | 3 - 7 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.825 ms | 2 - 6 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Qualcomm® QCS8450 | 7.206 ms | 4 - 223 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 6.1 ms | 3 - 7 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 6.105 ms | 3 - 3 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 39.408 ms | 0 - 197 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 7.552 ms | 4 - 202 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 2.68 ms | 1 - 89 MB | NPU |
| YOLOv11-Pose | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 7.552 ms | 4 - 202 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.27 ms | 5 - 194 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.838 ms | 5 - 178 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Snapdragon® X2 Elite | 3.148 ms | 5 - 5 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Snapdragon® X Elite | 5.899 ms | 5 - 5 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 3.784 ms | 0 - 214 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 9.046 ms | 5 - 207 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 6.359 ms | 5 - 14 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.369 ms | 5 - 7 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Qualcomm® QCS8450 | 9.046 ms | 5 - 207 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 6.593 ms | 5 - 13 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 5.899 ms | 5 - 5 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 2.838 ms | 5 - 178 MB | NPU |
| YOLOv11-Pose | QNN_DLC | float | Qualcomm® SA8295P | 9.402 ms | 0 - 178 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 1.875 ms | 2 - 193 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.197 ms | 0 - 174 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 2.341 ms | 2 - 2 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Snapdragon® X Elite | 5.186 ms | 2 - 2 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.124 ms | 2 - 200 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 4.452 ms | 2 - 8 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 4.679 ms | 2 - 81 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Qualcomm® SA8775P | 5.32 ms | 1 - 180 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Qualcomm® SA8650P | 5.32 ms | 1 - 180 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Qualcomm® SA8255P | 5.32 ms | 1 - 180 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 5.076 ms | 1 - 6 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 5.186 ms | 2 - 2 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 24.788 ms | 2 - 180 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 5.156 ms | 2 - 180 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 2.197 ms | 0 - 174 MB | NPU |
| YOLOv11-Pose | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 5.156 ms | 2 - 180 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.175 ms | 0 - 94 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.601 ms | 0 - 85 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 3.527 ms | 0 - 117 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 9.199 ms | 2 - 103 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 6.068 ms | 2 - 19 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.072 ms | 2 - 3 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Qualcomm® SA8775P | 6.558 ms | 2 - 83 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Qualcomm® SA8650P | 6.558 ms | 2 - 83 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Qualcomm® SA8255P | 6.558 ms | 2 - 83 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Qualcomm® QCS8450 | 9.199 ms | 2 - 103 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 6.33 ms | 0 - 16 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 2.601 ms | 0 - 85 MB | NPU |
| YOLOv11-Pose | TFLITE | float | Qualcomm® SA7255P | 50.28 ms | 0 - 27 MB | GPU |
| YOLOv11-Pose | TFLITE | float | Qualcomm® SA8295P | 9.326 ms | 2 - 71 MB | NPU |
License
- The license for the original implementation of YOLOv11-Pose can be found here.
References
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.
