BEVDet: Optimized for Qualcomm Devices

BEVDet is a machine learning model for generating a birds eye view represenation from the sensors(cameras) mounted on a vehicle.

This is based on the implementation of BEVDet 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	ONNX Runtime 1.25.0	Download
TFLITE	float	Universal		Download

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

Model Details

Model Type: Model_use_case.driver_assistance

Model Stats:

• Model checkpoint: bevdet-r50.pth
• Input resolution: 1 x 6 x 3 x 256 x 704
• Number of parameters: 44M
• Model size: 171 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
BEVDet	ONNX	float	Snapdragon® X2 Elite	570.115 ms	734 - 734 MB	CPU
BEVDet	ONNX	float	Snapdragon® X Elite	3019.739 ms	466 - 466 MB	CPU
BEVDet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	2358.617 ms	211 - 221 MB	CPU
BEVDet	ONNX	float	Snapdragon® 8 Gen 1 Mobile	3077.073 ms	204 - 219 MB	CPU
BEVDet	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	2505.446 ms	187 - 189 MB	CPU
BEVDet	ONNX	float	Qualcomm® QCS8450	3077.073 ms	204 - 219 MB	CPU
BEVDet	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	1306.431 ms	217 - 224 MB	CPU
BEVDet	ONNX	float	Snapdragon® 8 Elite Mobile	1457.629 ms	239 - 246 MB	CPU
BEVDet	ONNX	float	Qualcomm® Dragonwing™ Q-8750	1457.629 ms	239 - 246 MB	CPU
BEVDet	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	3019.739 ms	466 - 466 MB	CPU
BEVDet	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	1506.065 ms	241 - 255 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Snapdragon® X2 Elite	724.814 ms	712 - 712 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Snapdragon® X Elite	4253.027 ms	709 - 709 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Snapdragon® 8 Gen 3 Mobile	2437.909 ms	303 - 317 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Snapdragon® 8 Gen 1 Mobile	3113.118 ms	418 - 440 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	2818.254 ms	397 - 400 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Qualcomm® QCS8450	3113.118 ms	418 - 440 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Snapdragon® 8 Elite Gen 5 Mobile	1423.65 ms	288 - 299 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Qualcomm® Dragonwing™ IQ-9075	1869.81 ms	423 - 433 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Qualcomm® Dragonwing™ IQ-X7181	4253.027 ms	709 - 709 MB	CPU
BEVDet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	1764.127 ms	125 - 141 MB	CPU
BEVDet	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	2779.243 ms	125 - 142 MB	CPU
BEVDet	TFLITE	float	Qualcomm® QCS8275	3163.413 ms	128 - 137 MB	CPU
BEVDet	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	2124.109 ms	128 - 132 MB	CPU
BEVDet	TFLITE	float	Qualcomm® SA8775P	2477.149 ms	128 - 138 MB	CPU
BEVDet	TFLITE	float	Qualcomm® SA8650P	2477.149 ms	128 - 138 MB	CPU
BEVDet	TFLITE	float	Qualcomm® SA8255P	2477.149 ms	128 - 138 MB	CPU
BEVDet	TFLITE	float	Qualcomm® QCS8450	2779.243 ms	125 - 142 MB	CPU
BEVDet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	389.525 ms	127 - 140 MB	CPU
BEVDet	TFLITE	float	Snapdragon® 8 Elite Mobile	1240.134 ms	78 - 88 MB	CPU
BEVDet	TFLITE	float	Qualcomm® SA7255P	3163.413 ms	128 - 137 MB	CPU
BEVDet	TFLITE	float	Qualcomm® SA8295P	1845.776 ms	128 - 138 MB	CPU
BEVDet	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	1240.134 ms	78 - 88 MB	CPU
BEVDet	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	2392.932 ms	126 - 1330 MB	CPU

License

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

References

• BEVDet: High-Performance Multi-Camera 3D Object Detection in Bird-Eye-View
• 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.