BEVFusion: Optimized for Qualcomm Devices

BeVFusion 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 BEVFusion 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
PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	QAIRT 2.37, ONNX Runtime 1.23.0	Download
PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	QAIRT 2.37, ONNX Runtime 1.23.0	Download
PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	QAIRT 2.37, ONNX Runtime 1.23.0	Download
PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	QAIRT 2.37, ONNX Runtime 1.23.0	Download
PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	QAIRT 2.37, ONNX Runtime 1.23.0	Download
QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	QAIRT 2.42	Download
QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	QAIRT 2.42	Download
QNN_CONTEXT_BINARY	float	Qualcomm® SA8775P	QAIRT 2.42	Download
QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite For Galaxy Mobile	QAIRT 2.42	Download
QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	QAIRT 2.42	Download

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

Model Details

Model Type: Model_use_case.driver_assistance

Model Stats:

• Model checkpoint: camera-only-det.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
BEVFusionDecoder	PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	13.201 ms	24 - 24 MB	NPU
BEVFusionDecoder	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	9.916 ms	17 - 25 MB	NPU
BEVFusionDecoder	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	24.227 ms	21 - 24 MB	NPU
BEVFusionDecoder	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	7.889 ms	12 - 19 MB	NPU
BEVFusionDecoder	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	6.201 ms	17 - 28 MB	NPU
BEVFusionDecoder	QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	13.08 ms	5 - 5 MB	NPU
BEVFusionDecoder	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	9.848 ms	5 - 12 MB	NPU
BEVFusionDecoder	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite For Galaxy Mobile	7.597 ms	5 - 14 MB	NPU
BEVFusionDecoder	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	5.729 ms	5 - 15 MB	NPU
BEVFusionEncoder1	PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	816.0 ms	97 - 97 MB	NPU
BEVFusionEncoder1	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	618.339 ms	33 - 45 MB	NPU
BEVFusionEncoder1	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	835.545 ms	12 - 28 MB	NPU
BEVFusionEncoder1	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	508.583 ms	21 - 33 MB	NPU
BEVFusionEncoder1	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	432.064 ms	50 - 61 MB	NPU
BEVFusionEncoder1	QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	683.073 ms	12 - 12 MB	NPU
BEVFusionEncoder1	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	521.79 ms	13 - 20 MB	NPU
BEVFusionEncoder1	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite For Galaxy Mobile	420.381 ms	0 - 10 MB	NPU
BEVFusionEncoder1	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	350.55 ms	12 - 21 MB	NPU
BEVFusionEncoder2	PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	3251.893 ms	1058 - 1058 MB	NPU
BEVFusionEncoder2	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	2619.194 ms	516 - 524 MB	NPU
BEVFusionEncoder2	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	3309.274 ms	476 - 478 MB	NPU
BEVFusionEncoder2	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	2390.471 ms	546 - 558 MB	NPU
BEVFusionEncoder2	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	2189.505 ms	509 - 520 MB	NPU
BEVFusionEncoder2	QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	3419.442 ms	17 - 17 MB	NPU
BEVFusionEncoder2	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	2697.193 ms	17 - 25 MB	NPU
BEVFusionEncoder2	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite For Galaxy Mobile	2439.831 ms	3 - 16 MB	NPU
BEVFusionEncoder2	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	2186.323 ms	0 - 10 MB	NPU
BEVFusionEncoder3	PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	513.549 ms	610 - 610 MB	NPU
BEVFusionEncoder3	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	457.232 ms	578 - 590 MB	NPU
BEVFusionEncoder3	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	538.716 ms	608 - 1220 MB	NPU
BEVFusionEncoder3	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	374.412 ms	596 - 609 MB	NPU
BEVFusionEncoder3	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	335.284 ms	509 - 520 MB	NPU
BEVFusionEncoder3	QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	700.145 ms	610 - 610 MB	NPU
BEVFusionEncoder3	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	591.369 ms	609 - 621 MB	NPU
BEVFusionEncoder3	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite For Galaxy Mobile	499.751 ms	609 - 623 MB	NPU
BEVFusionEncoder3	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	369.428 ms	609 - 620 MB	NPU
BEVFusionEncoder4	PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	12.081 ms	19 - 19 MB	NPU
BEVFusionEncoder4	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	9.452 ms	32 - 39 MB	NPU
BEVFusionEncoder4	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	15.331 ms	18 - 39 MB	NPU
BEVFusionEncoder4	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	8.075 ms	12 - 19 MB	NPU
BEVFusionEncoder4	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	7.115 ms	30 - 40 MB	NPU
BEVFusionEncoder4	QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	11.877 ms	19 - 19 MB	NPU
BEVFusionEncoder4	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	9.138 ms	18 - 25 MB	NPU
BEVFusionEncoder4	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite For Galaxy Mobile	7.783 ms	18 - 27 MB	NPU
BEVFusionEncoder4	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	6.77 ms	18 - 28 MB	NPU

License

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

References

• BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird's-Eye View Representation
• 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.