DETR-ResNet101: 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-ResNet101 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 2.45, ONNX Runtime 1.27.1	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit DETR-ResNet101 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-ResNet101 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.object_detection

Model Stats:

• Model checkpoint: ResNet101
• Input resolution: 480x480
• Number of parameters: 60.3M
• Model size (float): 230 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
DETR-ResNet101	ONNX	float	Snapdragon® X2 Elite	12.755 ms	5 - 5 MB	NPU
DETR-ResNet101	ONNX	float	Snapdragon® X Elite	24.657 ms	114 - 114 MB	NPU
DETR-ResNet101	ONNX	float	Snapdragon® 8 Gen 3 Mobile	18.486 ms	1 - 467 MB	NPU
DETR-ResNet101	ONNX	float	Snapdragon® 8 Gen 1 Mobile	49.1 ms	1 - 363 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	24.665 ms	0 - 126 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® QCS8450	49.1 ms	1 - 363 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	38.652 ms	5 - 12 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	24.657 ms	114 - 114 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® Dragonwing™ Q-8750	14.587 ms	3 - 327 MB	NPU
DETR-ResNet101	ONNX	float	Snapdragon® 8 Elite Mobile	14.587 ms	3 - 327 MB	NPU
DETR-ResNet101	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	11.81 ms	2 - 339 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® X2 Elite	12.733 ms	5 - 5 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® X Elite	27.088 ms	5 - 5 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	19.775 ms	2 - 435 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	55.528 ms	5 - 342 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8275	138.081 ms	2 - 319 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	26.38 ms	5 - 7 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® SA8775P	41.139 ms	1 - 318 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® SA8650P	41.139 ms	1 - 318 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® SA8255P	41.139 ms	1 - 318 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® QCS8450	55.528 ms	5 - 342 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	41.759 ms	7 - 13 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	27.088 ms	5 - 5 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	15.296 ms	0 - 323 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® SA7255P	138.081 ms	2 - 319 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® SA8295P	43.217 ms	0 - 233 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® 8 Elite Mobile	15.296 ms	0 - 323 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	11.067 ms	5 - 345 MB	NPU
DETR-ResNet101	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	19.866 ms	0 - 496 MB	NPU
DETR-ResNet101	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	54.82 ms	0 - 390 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® Dragonwing™ QCS8275	138.036 ms	0 - 367 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	27.065 ms	0 - 3 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® SA8775P	41.234 ms	0 - 367 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® SA8650P	41.234 ms	0 - 367 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® SA8255P	41.234 ms	0 - 367 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® QCS8450	54.82 ms	0 - 390 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	41.039 ms	0 - 125 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	15.075 ms	0 - 372 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® SA7255P	138.036 ms	0 - 367 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® SA8295P	43.351 ms	0 - 275 MB	NPU
DETR-ResNet101	TFLITE	float	Snapdragon® 8 Elite Mobile	15.075 ms	0 - 372 MB	NPU
DETR-ResNet101	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	11.063 ms	0 - 393 MB	NPU

License

• The license for the original implementation of DETR-ResNet101 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.