PointNet: Optimized for Qualcomm Devices

PointNet is a pioneering neural network architecture designed to directly consume unordered point cloud data for tasks such as classification and segmentation. It learns spatial features from raw 3D points without requiring voxelization or image projections.

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 2.42, ONNX Runtime 1.24.3	Download
QNN_DLC	float	Universal	QAIRT 2.43	Download
TFLITE	float	Universal	QAIRT 2.43, TFLite 2.19.1	Download

For more device-specific assets and performance metrics, visit PointNet 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 PointNet on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

• Model checkpoint: save
• Input resolution: 1x3x1024
• Model size: 13.2 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
PointNet	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	0.319 ms	0 - 29 MB	NPU
PointNet	ONNX	float	Snapdragon® X2 Elite	0.314 ms	7 - 7 MB	NPU
PointNet	ONNX	float	Snapdragon® X Elite	0.784 ms	7 - 7 MB	NPU
PointNet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	0.405 ms	0 - 42 MB	NPU
PointNet	ONNX	float	Qualcomm® QCS8550 (Proxy)	0.648 ms	0 - 8 MB	NPU
PointNet	ONNX	float	Qualcomm® QCS9075	0.826 ms	0 - 3 MB	NPU
PointNet	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	0.435 ms	0 - 25 MB	NPU
PointNet	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	0.32 ms	0 - 30 MB	NPU
PointNet	QNN_DLC	float	Snapdragon® X2 Elite	0.426 ms	1 - 1 MB	NPU
PointNet	QNN_DLC	float	Snapdragon® X Elite	0.801 ms	0 - 0 MB	NPU
PointNet	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	0.414 ms	0 - 43 MB	NPU
PointNet	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	1.857 ms	0 - 27 MB	NPU
PointNet	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	0.663 ms	0 - 6 MB	NPU
PointNet	QNN_DLC	float	Qualcomm® QCS9075	0.798 ms	2 - 4 MB	NPU
PointNet	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	0.867 ms	0 - 44 MB	NPU
PointNet	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	0.477 ms	0 - 26 MB	NPU
PointNet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	0.315 ms	0 - 30 MB	NPU
PointNet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	0.412 ms	0 - 44 MB	NPU
PointNet	TFLITE	float	Qualcomm® QCS8275 (Proxy)	1.9 ms	0 - 28 MB	NPU
PointNet	TFLITE	float	Qualcomm® QCS8550 (Proxy)	0.663 ms	0 - 2 MB	NPU
PointNet	TFLITE	float	Qualcomm® QCS9075	0.804 ms	0 - 9 MB	NPU
PointNet	TFLITE	float	Qualcomm® QCS8450 (Proxy)	0.863 ms	0 - 45 MB	NPU
PointNet	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	0.485 ms	0 - 27 MB	NPU

License

• The license for the original implementation of PointNet can be found here.

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.