ResNeXt50: Optimized for Qualcomm Devices

ResNeXt50 is a machine learning model that can classify images from the Imagenet dataset. It can also be used as a backbone in building more complex models for specific use cases.

This is based on the implementation of ResNeXt50 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	w8a8	Universal	QAIRT 2.42, ONNX Runtime 1.25.0	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
QNN_DLC	w8a8	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download
TFLITE	w8a8	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit ResNeXt50 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 ResNeXt50 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

• Model checkpoint: Imagenet
• Input resolution: 224x224
• Number of parameters: 25.0M
• Model size (float): 95.4 MB
• Model size (w8a8): 24.8 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
ResNeXt50	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	1.142 ms	1 - 77 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® X2 Elite	1.097 ms	212 - 212 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® X Elite	2.189 ms	149 - 149 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® 8 Gen 3 Mobile	1.655 ms	1 - 142 MB	NPU
ResNeXt50	ONNX	float	Qualcomm® QCS8550 (Proxy)	2.295 ms	0 - 69 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	1.386 ms	0 - 78 MB	NPU
ResNeXt50	ONNX	float	Qualcomm® QCS9075	3.457 ms	1 - 46 MB	NPU
ResNeXt50	ONNX	float	Qualcomm® QCS8750	1.386 ms	0 - 78 MB	NPU
ResNeXt50	ONNX	float	Qualcomm® QCS7181	2.189 ms	149 - 149 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.596 ms	0 - 78 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® X Elite	1.101 ms	181 - 181 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.83 ms	0 - 108 MB	NPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS6490	47.211 ms	6 - 27 MB	CPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	1.128 ms	0 - 16 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	20.507 ms	4 - 13 MB	CPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCM6690	26.024 ms	3 - 12 MB	CPU
ResNeXt50	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.697 ms	0 - 76 MB	NPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS9075	1.242 ms	0 - 45 MB	NPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS7790	20.507 ms	4 - 13 MB	CPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS8750	0.697 ms	0 - 76 MB	NPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS7181	1.101 ms	181 - 181 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	1.171 ms	1 - 74 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® X2 Elite	1.393 ms	1 - 1 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® X Elite	2.675 ms	1 - 1 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	1.795 ms	0 - 135 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS8275	12.041 ms	1 - 72 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	2.571 ms	1 - 170 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA8775P	3.815 ms	1 - 73 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA8650P	3.815 ms	1 - 73 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA8255P	3.815 ms	1 - 73 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	4.034 ms	0 - 113 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA8295P	4.118 ms	0 - 52 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	1.46 ms	0 - 73 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA7255P	12.041 ms	1 - 72 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS9075	3.637 ms	1 - 3 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS8750	1.46 ms	0 - 73 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS7181	2.675 ms	1 - 1 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.538 ms	0 - 77 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.67 ms	0 - 0 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® X Elite	1.246 ms	0 - 0 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.798 ms	0 - 102 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS6490	3.122 ms	0 - 2 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS8275	2.446 ms	0 - 73 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	1.096 ms	0 - 2 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA8775P	1.453 ms	0 - 74 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA8650P	1.453 ms	0 - 74 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA8255P	1.453 ms	0 - 74 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA7255P	2.446 ms	0 - 73 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	1.374 ms	0 - 77 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA8295P	1.766 ms	0 - 71 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCM6690	7.641 ms	0 - 196 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.635 ms	0 - 74 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS9075	1.209 ms	0 - 2 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	1.499 ms	0 - 101 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS7790	1.374 ms	0 - 77 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS8750	0.635 ms	0 - 74 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS7181	1.246 ms	0 - 0 MB	NPU
ResNeXt50	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	1.177 ms	0 - 115 MB	NPU
ResNeXt50	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	1.768 ms	0 - 176 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS8275	12.03 ms	0 - 113 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS8550 (Proxy)	2.475 ms	0 - 2 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA8775P	3.836 ms	0 - 113 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA8650P	3.836 ms	0 - 113 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA8255P	3.836 ms	0 - 113 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS8450 (Proxy)	4.007 ms	0 - 153 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA8295P	4.146 ms	0 - 86 MB	NPU
ResNeXt50	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	1.429 ms	0 - 113 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA7255P	12.03 ms	0 - 113 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS9075	3.692 ms	0 - 52 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS8750	1.429 ms	0 - 113 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.485 ms	0 - 73 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.652 ms	0 - 100 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS6490	2.846 ms	0 - 27 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS8275	2.305 ms	0 - 70 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	0.917 ms	0 - 2 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA8775P	1.263 ms	0 - 72 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA8650P	1.263 ms	0 - 72 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA8255P	1.263 ms	0 - 72 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA7255P	2.305 ms	0 - 70 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	1.212 ms	0 - 70 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA8295P	1.567 ms	0 - 64 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCM6690	7.363 ms	0 - 198 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.556 ms	0 - 63 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS9075	0.993 ms	0 - 27 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	1.282 ms	0 - 100 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS7790	1.212 ms	0 - 70 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS8750	0.556 ms	0 - 63 MB	NPU

License

• The license for the original implementation of ResNeXt50 can be found here.

References

• Aggregated Residual Transformations for Deep Neural Networks
• 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.