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
ONNX	w8a16_mixed_fp16	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® 8 Elite Gen 5 Mobile	1466.13 ms	251 - 262 MB	CPU
BEVDet	ONNX	float	Snapdragon® X2 Elite	602.216 ms	734 - 734 MB	CPU
BEVDet	ONNX	float	Snapdragon® X Elite	2396.069 ms	467 - 467 MB	CPU
BEVDet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	2117.801 ms	217 - 228 MB	CPU
BEVDet	ONNX	float	Qualcomm® QCS8550 (Proxy)	2431.029 ms	188 - 190 MB	CPU
BEVDet	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	1508.308 ms	237 - 245 MB	CPU
BEVDet	ONNX	float	Qualcomm® QCS9075	1502.705 ms	241 - 255 MB	CPU
BEVDet	ONNX	float	Qualcomm® QCS8750	1508.308 ms	237 - 245 MB	CPU
BEVDet	ONNX	float	Qualcomm® QCS7181	2396.069 ms	467 - 467 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Snapdragon® 8 Elite Gen 5 Mobile	1772.158 ms	267 - 281 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Snapdragon® X2 Elite	821.804 ms	1239 - 1239 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Snapdragon® 8 Gen 3 Mobile	2340.571 ms	361 - 372 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Snapdragon® 8 Elite For Galaxy Mobile	1538.833 ms	322 - 333 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Qualcomm® QCS9075	1843.327 ms	422 - 431 MB	CPU
BEVDet	ONNX	w8a16_mixed_fp16	Qualcomm® QCS8750	1538.833 ms	322 - 333 MB	CPU
BEVDet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	384.362 ms	87 - 98 MB	CPU
BEVDet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	1647.156 ms	121 - 136 MB	CPU
BEVDet	TFLITE	float	Qualcomm® QCS8275	3176.258 ms	128 - 137 MB	CPU
BEVDet	TFLITE	float	Qualcomm® QCS8550 (Proxy)	2039.323 ms	2 - 253 MB	CPU
BEVDet	TFLITE	float	Qualcomm® SA8775P	2497.786 ms	128 - 138 MB	CPU
BEVDet	TFLITE	float	Qualcomm® SA8650P	2497.786 ms	128 - 138 MB	CPU
BEVDet	TFLITE	float	Qualcomm® SA8255P	2497.786 ms	128 - 138 MB	CPU
BEVDet	TFLITE	float	Qualcomm® QCS8450 (Proxy)	2697.925 ms	125 - 141 MB	CPU
BEVDet	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	1199.523 ms	76 - 85 MB	CPU
BEVDet	TFLITE	float	Qualcomm® QCS9075	2393.421 ms	127 - 1330 MB	CPU
BEVDet	TFLITE	float	Qualcomm® SA7255P	3176.258 ms	128 - 137 MB	CPU
BEVDet	TFLITE	float	Qualcomm® SA8295P	2068.392 ms	128 - 138 MB	CPU
BEVDet	TFLITE	float	Qualcomm® QCS8750	1199.523 ms	76 - 85 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.