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
ONNX	w8a16_mixed_int16	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:

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

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
DETR-ResNet101	ONNX	float	Snapdragon® X2 Elite	12.58 ms	5 - 5 MB	NPU
DETR-ResNet101	ONNX	float	Snapdragon® X Elite	24.784 ms	115 - 115 MB	NPU
DETR-ResNet101	ONNX	float	Snapdragon® 8 Gen 3 Mobile	18.406 ms	3 - 455 MB	NPU
DETR-ResNet101	ONNX	float	Snapdragon® 8 Gen 1 Mobile	49.921 ms	4 - 366 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	42.754 ms	5 - 13 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	24.889 ms	0 - 163 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® QCS8450	49.921 ms	4 - 366 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	38.737 ms	5 - 12 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	24.784 ms	115 - 115 MB	NPU
DETR-ResNet101	ONNX	float	Qualcomm® Dragonwing™ Q-8750	14.655 ms	0 - 326 MB	NPU
DETR-ResNet101	ONNX	float	Snapdragon® 8 Elite Mobile	14.655 ms	0 - 326 MB	NPU
DETR-ResNet101	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	11.808 ms	3 - 342 MB	NPU
DETR-ResNet101	ONNX	w8a16_mixed_int16	Snapdragon® 8 Gen 3 Mobile	49.597 ms	20 - 470 MB	NPU
DETR-ResNet101	ONNX	w8a16_mixed_int16	Qualcomm® Dragonwing™ Q-8750	42.823 ms	23 - 324 MB	NPU
DETR-ResNet101	ONNX	w8a16_mixed_int16	Snapdragon® 8 Elite Mobile	42.823 ms	23 - 324 MB	NPU
DETR-ResNet101	ONNX	w8a16_mixed_int16	Snapdragon® 8 Elite Gen 5 Mobile	34.906 ms	21 - 368 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® X2 Elite	12.767 ms	5 - 5 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® X Elite	27.085 ms	5 - 5 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	19.745 ms	5 - 440 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	54.916 ms	4 - 342 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	44.136 ms	5 - 12 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	137.997 ms	2 - 319 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	26.61 ms	5 - 8 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® SA8775P	41.171 ms	1 - 318 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® SA8650P	41.171 ms	1 - 318 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® SA8255P	41.171 ms	1 - 318 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® QCS8450	54.916 ms	4 - 342 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	41.91 ms	5 - 11 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	27.085 ms	5 - 5 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	15.193 ms	0 - 322 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® SA7255P	137.997 ms	2 - 319 MB	NPU
DETR-ResNet101	QNN_DLC	float	Qualcomm® SA8295P	43.234 ms	0 - 233 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® 8 Elite Mobile	15.193 ms	0 - 322 MB	NPU
DETR-ResNet101	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	11.049 ms	5 - 345 MB	NPU
DETR-ResNet101	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	19.698 ms	0 - 491 MB	NPU
DETR-ResNet101	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	56.482 ms	0 - 390 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	44.081 ms	0 - 126 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	138.173 ms	0 - 367 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	26.668 ms	0 - 3 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® SA8775P	41.216 ms	0 - 367 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® SA8650P	41.216 ms	0 - 367 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® SA8255P	41.216 ms	0 - 367 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® QCS8450	56.482 ms	0 - 390 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	41.894 ms	0 - 125 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	15.269 ms	0 - 372 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® SA7255P	138.173 ms	0 - 367 MB	NPU
DETR-ResNet101	TFLITE	float	Qualcomm® SA8295P	43.358 ms	0 - 274 MB	NPU
DETR-ResNet101	TFLITE	float	Snapdragon® 8 Elite Mobile	15.269 ms	0 - 372 MB	NPU
DETR-ResNet101	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	11.073 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.