RangeNet-Plus-Plus: Optimized for Qualcomm Devices
RangeNet-Plus-Plus (also stylized as RangeNet++) projects a LiDAR point cloud onto a 5-channel range image (depth, x, y, z, intensity) and applies a DarkNet-53 encoder with a decoder head to predict per-point semantic class labels in real time.
This is based on the implementation of RangeNet-Plus-Plus 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
There are two ways to deploy this model on your device:
Option 1: Download Pre-Exported Models
Below are pre-exported model assets ready for deployment.
| Runtime | Precision | Chipset | SDK Versions | Download |
|---|---|---|---|---|
| ONNX | float | Universal | QAIRT 2.50, ONNX Runtime 1.27.1 | Download |
| ONNX | w8a16 | Universal | QAIRT 2.50, ONNX Runtime 1.27.1 | Download |
| QNN_DLC | float | Universal | QAIRT 2.50 | Download |
| QNN_DLC | w8a16 | Universal | QAIRT 2.50 | Download |
| TFLITE | float | Universal | QAIRT 2.50 | Download |
For more device-specific assets and performance metrics, visit RangeNet-Plus-Plus on Qualcomm® AI Hub.
Option 2: Export with Custom Configurations
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
This option is ideal if you need to customize the model beyond the default configuration provided here.
See our repository for RangeNet-Plus-Plus on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.driver_assistance
Model Stats:
- Backbone: DarkNet-53
- Input channels: 5
- Input resolution: 64x2048
- Model checkpoint: darknet53_rangenet++
- Number of output classes: 20
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| RangeNet-Plus-Plus | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 43.136 ms | 24 - 360 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 59.67 ms | 21 - 332 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Snapdragon® X2 Elite | 49.72 ms | 33 - 33 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Snapdragon® X Elite | 98.486 ms | 101 - 101 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 76.691 ms | 0 - 450 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 222.22 ms | 1 - 446 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 176.55 ms | 21 - 28 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 102.453 ms | 0 - 112 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Qualcomm® QCS8450 | 222.22 ms | 1 - 446 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 150.417 ms | 21 - 27 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 98.486 ms | 101 - 101 MB | NPU |
| RangeNet-Plus-Plus | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 59.67 ms | 21 - 332 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 28.087 ms | 11 - 353 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 38.797 ms | 14 - 346 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Snapdragon® X2 Elite | 32.071 ms | 18 - 18 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Snapdragon® X Elite | 60.977 ms | 52 - 52 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 48.393 ms | 0 - 460 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 95.074 ms | 0 - 570 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 408.191 ms | 26 - 31 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 62.931 ms | 11 - 16 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 61.608 ms | 0 - 58 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Qualcomm® QCS8450 | 95.074 ms | 0 - 570 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 58.107 ms | 14 - 18 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 60.977 ms | 52 - 52 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 1592.222 ms | 0 - 726 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 118.454 ms | 28 - 476 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 38.797 ms | 14 - 346 MB | NPU |
| RangeNet-Plus-Plus | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 118.454 ms | 28 - 476 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 44.729 ms | 3 - 332 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 60.768 ms | 0 - 304 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Snapdragon® X2 Elite | 49.596 ms | 3 - 3 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Snapdragon® X Elite | 98.956 ms | 3 - 3 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 77.224 ms | 3 - 451 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 181.194 ms | 3 - 439 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 176.544 ms | 3 - 18 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 100.72 ms | 3 - 5 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Qualcomm® SA8650P | 152.862 ms | 1 - 334 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Qualcomm® SA8255P | 152.862 ms | 1 - 334 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Qualcomm® QCS8450 | 181.194 ms | 3 - 439 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 150.135 ms | 3 - 17 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 98.956 ms | 3 - 3 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 60.768 ms | 0 - 304 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Qualcomm® SA7255P | 592.921 ms | 1 - 319 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | float | Qualcomm® SA8295P | 166.76 ms | 0 - 312 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 26.981 ms | 1 - 330 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 40.943 ms | 1 - 326 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 32.084 ms | 1 - 1 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Snapdragon® X Elite | 61.492 ms | 1 - 1 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 48.449 ms | 1 - 452 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Snapdragon® 8 Gen 1 Mobile | 98.146 ms | 1 - 563 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 403.134 ms | 3 - 11 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 63.646 ms | 1 - 10 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 61.311 ms | 1 - 3 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® SA8650P | 62.375 ms | 1 - 326 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® SA8255P | 62.375 ms | 1 - 326 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® QCS8450 | 98.146 ms | 1 - 563 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 57.543 ms | 1 - 9 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 61.492 ms | 1 - 1 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 1598.647 ms | 2 - 704 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 115.095 ms | 1 - 437 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 40.943 ms | 1 - 326 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® SA7255P | 186.577 ms | 2 - 318 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Qualcomm® SA8295P | 101.809 ms | 2 - 412 MB | NPU |
| RangeNet-Plus-Plus | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 115.095 ms | 1 - 437 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 43.631 ms | 0 - 307 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 61.186 ms | 12 - 296 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 78.205 ms | 12 - 522 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 186.706 ms | 13 - 504 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 178.935 ms | 13 - 134 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 100.128 ms | 5 - 401 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Qualcomm® SA8650P | 153.857 ms | 13 - 322 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Qualcomm® SA8255P | 153.857 ms | 13 - 322 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Qualcomm® QCS8450 | 186.706 ms | 13 - 504 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 153.321 ms | 12 - 132 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 61.186 ms | 12 - 296 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Qualcomm® SA7255P | 592.906 ms | 13 - 307 MB | NPU |
| RangeNet-Plus-Plus | TFLITE | float | Qualcomm® SA8295P | 167.763 ms | 13 - 301 MB | NPU |
License
- The license for the original implementation of RangeNet-Plus-Plus 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.
