Conditional-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 Conditional-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 1.25.0	Download
ONNX	w8a16_mixed_fp16	Universal	QAIRT 2.42, ONNX Runtime 1.25.0	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
QNN_DLC	w8a16_mixed_fp16	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit Conditional-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 Conditional-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: 43.6M
• Model size (float): 166 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
Conditional-DETR-ResNet50	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	8.314 ms	5 - 359 MB	NPU
Conditional-DETR-ResNet50	ONNX	float	Snapdragon® X2 Elite	9.356 ms	207 - 207 MB	NPU
Conditional-DETR-ResNet50	ONNX	float	Snapdragon® X Elite	20.272 ms	155 - 155 MB	NPU
Conditional-DETR-ResNet50	ONNX	float	Snapdragon® 8 Gen 3 Mobile	14.933 ms	5 - 465 MB	NPU
Conditional-DETR-ResNet50	ONNX	float	Qualcomm® QCS8550 (Proxy)	20.289 ms	0 - 95 MB	NPU
Conditional-DETR-ResNet50	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	11.075 ms	0 - 330 MB	NPU
Conditional-DETR-ResNet50	ONNX	float	Qualcomm® QCS9075	30.487 ms	5 - 50 MB	NPU
Conditional-DETR-ResNet50	ONNX	float	Qualcomm® QCS8750	11.075 ms	0 - 330 MB	NPU
Conditional-DETR-ResNet50	ONNX	float	Qualcomm® QCS7181	20.272 ms	155 - 155 MB	NPU
Conditional-DETR-ResNet50	ONNX	w8a16_mixed_fp16	Snapdragon® 8 Elite Gen 5 Mobile	25.573 ms	13 - 420 MB	NPU
Conditional-DETR-ResNet50	ONNX	w8a16_mixed_fp16	Snapdragon® 8 Gen 3 Mobile	45.233 ms	17 - 521 MB	NPU
Conditional-DETR-ResNet50	ONNX	w8a16_mixed_fp16	Snapdragon® 8 Elite For Galaxy Mobile	33.978 ms	6 - 388 MB	NPU
Conditional-DETR-ResNet50	ONNX	w8a16_mixed_fp16	Qualcomm® QCS8750	33.978 ms	6 - 388 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	8.864 ms	5 - 332 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Snapdragon® X2 Elite	9.998 ms	5 - 5 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Snapdragon® X Elite	23.375 ms	5 - 5 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	17.104 ms	0 - 397 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® QCS8275	98.019 ms	1 - 309 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	23.045 ms	5 - 405 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® SA8775P	32.186 ms	2 - 309 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® SA8650P	32.186 ms	2 - 309 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® SA8255P	32.186 ms	2 - 309 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	44.951 ms	3 - 348 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® SA7255P	98.019 ms	1 - 309 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® SA8295P	34.293 ms	0 - 262 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	11.607 ms	0 - 316 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® QCS9075	36.588 ms	5 - 11 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® QCS8750	11.607 ms	0 - 316 MB	NPU
Conditional-DETR-ResNet50	QNN_DLC	float	Qualcomm® QCS7181	23.375 ms	5 - 5 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	9.224 ms	0 - 353 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	15.686 ms	0 - 435 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Qualcomm® QCS8275	93.293 ms	0 - 337 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Qualcomm® QCS8550 (Proxy)	22.049 ms	0 - 3 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Qualcomm® SA8775P	29.979 ms	0 - 368 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Qualcomm® SA8650P	29.979 ms	0 - 368 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Qualcomm® SA8255P	29.979 ms	0 - 368 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Qualcomm® QCS8450 (Proxy)	46.477 ms	0 - 376 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Qualcomm® SA7255P	93.293 ms	0 - 337 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Qualcomm® SA8295P	34.651 ms	0 - 290 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	11.927 ms	0 - 341 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Qualcomm® QCS9075	33.178 ms	0 - 93 MB	NPU
Conditional-DETR-ResNet50	TFLITE	float	Qualcomm® QCS8750	11.927 ms	0 - 341 MB	NPU

License

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

References

• Conditional {DETR} for Fast Training Convergence
• 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.