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® X2 Elite	QAIRT 2.45, ONNX Runtime 1.25.0	Download
PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	QAIRT 2.45, ONNX Runtime 1.25.0	Download
PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	QAIRT 2.45, ONNX Runtime 1.25.0	Download
PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 1 Mobile	QAIRT 2.45	Download
PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS8550 (Proxy)	QAIRT 2.45, ONNX Runtime 1.25.0	Download
PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Mobile	QAIRT 2.45, ONNX Runtime 1.25.0	Download
PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	QAIRT 2.45, ONNX Runtime 1.25.0	Download
PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	QAIRT 2.45, ONNX Runtime 1.25.0	Download
QNN_CONTEXT_BINARY	float	Snapdragon® X2 Elite	QAIRT 2.45	Download
QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	QAIRT 2.45	Download
QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	QAIRT 2.45	Download
QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 1 Mobile	QAIRT 2.45	Download
QNN_CONTEXT_BINARY	float	Qualcomm® QCS8550 (Proxy)	QAIRT 2.45	Download
QNN_CONTEXT_BINARY	float	Qualcomm® SA8775P	QAIRT 2.45	Download
QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Mobile	QAIRT 2.45	Download
QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	QAIRT 2.45	Download
QNN_CONTEXT_BINARY	float	Qualcomm® SA7255P	QAIRT 2.45	Download
QNN_CONTEXT_BINARY	float	Qualcomm® SA8295P	QAIRT 2.45	Download
QNN_CONTEXT_BINARY	float	Qualcomm® QCS9075	QAIRT 2.45	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
decoder	PRECOMPILED_QNN_ONNX	float	Snapdragon® X2 Elite	6.991 ms	22 - 22 MB	NPU
decoder	PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	12.919 ms	23 - 23 MB	NPU
decoder	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	9.786 ms	15 - 21 MB	NPU
decoder	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS8550 (Proxy)	12.835 ms	4 - 32 MB	NPU
decoder	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	5.74 ms	13 - 20 MB	NPU
decoder	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	24.148 ms	15 - 18 MB	NPU
decoder	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Mobile	7.767 ms	13 - 20 MB	NPU
decoder	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS8750	7.767 ms	13 - 20 MB	NPU
decoder	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS7181	12.919 ms	23 - 23 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Snapdragon® X2 Elite	7.529 ms	5 - 5 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	13.156 ms	5 - 5 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	9.782 ms	5 - 12 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Qualcomm® QCS8275	92.815 ms	0 - 9 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	5.734 ms	0 - 9 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Qualcomm® SA7255P	92.815 ms	0 - 9 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Qualcomm® QCS9075	24.351 ms	5 - 16 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Mobile	7.746 ms	0 - 13 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Qualcomm® SA8295P	26.057 ms	0 - 6 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Qualcomm® QCS8750	7.746 ms	0 - 13 MB	NPU
decoder	QNN_CONTEXT_BINARY	float	Qualcomm® QCS7181	13.156 ms	5 - 5 MB	NPU
encoder_1	PRECOMPILED_QNN_ONNX	float	Snapdragon® X2 Elite	387.006 ms	66 - 66 MB	NPU
encoder_1	PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	699.365 ms	97 - 97 MB	NPU
encoder_1	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	533.381 ms	33 - 45 MB	NPU
encoder_1	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS8550 (Proxy)	725.129 ms	29 - 31 MB	NPU
encoder_1	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	354.804 ms	37 - 51 MB	NPU
encoder_1	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	704.614 ms	12 - 27 MB	NPU
encoder_1	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Mobile	440.132 ms	39 - 51 MB	NPU
encoder_1	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS8750	440.132 ms	39 - 51 MB	NPU
encoder_1	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS7181	699.365 ms	97 - 97 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Snapdragon® X2 Elite	387.878 ms	12 - 12 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	682.814 ms	12 - 12 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	522.181 ms	15 - 22 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Qualcomm® QCS8275	1072.614 ms	1 - 10 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	354.969 ms	0 - 9 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Qualcomm® SA7255P	1072.614 ms	1 - 10 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Qualcomm® QCS9075	704.131 ms	12 - 43 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Mobile	437.61 ms	12 - 21 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Qualcomm® SA8295P	815.423 ms	0 - 5 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Qualcomm® QCS8750	437.61 ms	12 - 21 MB	NPU
encoder_1	QNN_CONTEXT_BINARY	float	Qualcomm® QCS7181	682.814 ms	12 - 12 MB	NPU
encoder_2	PRECOMPILED_QNN_ONNX	float	Snapdragon® X2 Elite	2334.502 ms	1058 - 1058 MB	NPU
encoder_2	PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	3499.347 ms	1058 - 1058 MB	NPU
encoder_2	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	2784.2 ms	577 - 583 MB	NPU
encoder_2	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS8550 (Proxy)	3474.913 ms	556 - 569 MB	NPU
encoder_2	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	2165.598 ms	443 - 450 MB	NPU
encoder_2	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	3427.657 ms	260 - 262 MB	NPU
encoder_2	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Mobile	2455.934 ms	408 - 419 MB	NPU
encoder_2	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS8750	2455.934 ms	408 - 419 MB	NPU
encoder_2	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS7181	3499.347 ms	1058 - 1058 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Snapdragon® X2 Elite	2345.717 ms	17 - 17 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	3426.33 ms	17 - 17 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	2803.181 ms	17 - 30 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Qualcomm® QCS8275	5457.449 ms	16 - 25 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	2155.55 ms	17 - 31 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Qualcomm® SA7255P	5457.449 ms	16 - 25 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Qualcomm® QCS9075	3421.482 ms	17 - 644 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Mobile	2430.895 ms	17 - 31 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Qualcomm® SA8295P	3651.109 ms	1 - 11 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Qualcomm® QCS8750	2430.895 ms	17 - 31 MB	NPU
encoder_2	QNN_CONTEXT_BINARY	float	Qualcomm® QCS7181	3426.33 ms	17 - 17 MB	NPU
encoder_3	PRECOMPILED_QNN_ONNX	float	Snapdragon® X2 Elite	7.866 ms	18 - 18 MB	NPU
encoder_3	PRECOMPILED_QNN_ONNX	float	Snapdragon® X Elite	11.701 ms	18 - 18 MB	NPU
encoder_3	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Gen 3 Mobile	9.207 ms	32 - 38 MB	NPU
encoder_3	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS8550 (Proxy)	12.604 ms	23 - 27 MB	NPU
encoder_3	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	6.765 ms	13 - 20 MB	NPU
encoder_3	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS9075	15.336 ms	18 - 39 MB	NPU
encoder_3	PRECOMPILED_QNN_ONNX	float	Snapdragon® 8 Elite Mobile	7.715 ms	14 - 25 MB	NPU
encoder_3	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS8750	7.715 ms	14 - 25 MB	NPU
encoder_3	PRECOMPILED_QNN_ONNX	float	Qualcomm® QCS7181	11.701 ms	18 - 18 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Snapdragon® X2 Elite	8.696 ms	18 - 18 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Snapdragon® X Elite	11.975 ms	19 - 19 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Gen 3 Mobile	9.303 ms	19 - 26 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Qualcomm® QCS8275	25.315 ms	18 - 27 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Gen 5 Mobile	6.784 ms	17 - 25 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Qualcomm® SA7255P	25.315 ms	18 - 27 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Qualcomm® QCS9075	15.218 ms	18 - 43 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Snapdragon® 8 Elite Mobile	7.724 ms	18 - 27 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Qualcomm® SA8295P	14.841 ms	1 - 6 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Qualcomm® QCS8750	7.724 ms	18 - 27 MB	NPU
encoder_3	QNN_CONTEXT_BINARY	float	Qualcomm® QCS7181	11.975 ms	19 - 19 MB	NPU

License

The license for the original implementation of BEVFusion 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.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

Other

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Paper for qualcomm/BEVFusion

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

Paper • 2205.13542 • Published May 26, 2022