DETR-ResNet50: Optimized for Qualcomm Devices

DETR is a machine learning model that can detect objects (trained on COCO dataset).

This is based on the implementation of DETR-ResNet50 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	QAIRT 2.42, ONNX Runtime 1.24.1	Download
QNN_DLC	float	Universal	QAIRT 2.43	Download
TFLITE	float	Universal	QAIRT 2.43, TFLite 2.17.0	Download

For more device-specific assets and performance metrics, visit DETR-ResNet50 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 DETR-ResNet50 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.object_detection

Model Stats:

• Model checkpoint: ResNet50
• Input resolution: 480x480
• Number of parameters: 41.4M
• Model size (float): 158 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
DETR-ResNet50	ONNX	float	Snapdragon® X2 Elite	8.368 ms	77 - 77 MB	NPU
DETR-ResNet50	ONNX	float	Snapdragon® X Elite	18.005 ms	77 - 77 MB	NPU
DETR-ResNet50	ONNX	float	Snapdragon® 8 Gen 3 Mobile	13.518 ms	1 - 397 MB	NPU
DETR-ResNet50	ONNX	float	Qualcomm® QCS8550 (Proxy)	17.884 ms	0 - 95 MB	NPU
DETR-ResNet50	ONNX	float	Qualcomm® QCS9075	28.427 ms	5 - 12 MB	NPU
DETR-ResNet50	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	9.864 ms	2 - 316 MB	NPU
DETR-ResNet50	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	7.452 ms	0 - 322 MB	NPU
DETR-ResNet50	QNN_DLC	float	Snapdragon® X2 Elite	9.019 ms	5 - 5 MB	NPU
DETR-ResNet50	QNN_DLC	float	Snapdragon® X Elite	20.668 ms	5 - 5 MB	NPU
DETR-ResNet50	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	14.943 ms	5 - 388 MB	NPU
DETR-ResNet50	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	91.478 ms	1 - 285 MB	NPU
DETR-ResNet50	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	20.564 ms	5 - 6 MB	NPU
DETR-ResNet50	QNN_DLC	float	Qualcomm® SA8775P	133.562 ms	1 - 284 MB	NPU
DETR-ResNet50	QNN_DLC	float	Qualcomm® QCS9075	30.754 ms	5 - 11 MB	NPU
DETR-ResNet50	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	43.242 ms	2 - 327 MB	NPU
DETR-ResNet50	QNN_DLC	float	Qualcomm® SA7255P	91.478 ms	1 - 285 MB	NPU
DETR-ResNet50	QNN_DLC	float	Qualcomm® SA8295P	31.396 ms	0 - 240 MB	NPU
DETR-ResNet50	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	10.284 ms	0 - 294 MB	NPU
DETR-ResNet50	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	8.038 ms	5 - 303 MB	NPU
DETR-ResNet50	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	13.895 ms	0 - 422 MB	NPU
DETR-ResNet50	TFLITE	float	Qualcomm® QCS8275 (Proxy)	85.59 ms	0 - 323 MB	NPU
DETR-ResNet50	TFLITE	float	Qualcomm® QCS8550 (Proxy)	19.485 ms	0 - 3 MB	NPU
DETR-ResNet50	TFLITE	float	Qualcomm® SA8775P	26.994 ms	0 - 382 MB	NPU
DETR-ResNet50	TFLITE	float	Qualcomm® QCS9075	30.505 ms	0 - 88 MB	NPU
DETR-ResNet50	TFLITE	float	Qualcomm® QCS8450 (Proxy)	44.292 ms	0 - 356 MB	NPU
DETR-ResNet50	TFLITE	float	Qualcomm® SA7255P	85.59 ms	0 - 323 MB	NPU
DETR-ResNet50	TFLITE	float	Qualcomm® SA8295P	31.076 ms	0 - 270 MB	NPU
DETR-ResNet50	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	10.275 ms	0 - 337 MB	NPU
DETR-ResNet50	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	7.997 ms	0 - 342 MB	NPU

License

• The license for the original implementation of DETR-ResNet50 can be found here.

References

• End-to-End Object Detection with Transformers
• 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.