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 1.24.3	Download
ONNX	w8a8	Universal	QAIRT 2.42, ONNX Runtime 1.24.3	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	1.144 ms	1 - 86 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® 8 Elite Mobile	1.379 ms	0 - 84 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® X2 Elite	1.095 ms	50 - 50 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® X Elite	2.424 ms	50 - 50 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® X Elite	2.424 ms	50 - 50 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® 8 Gen 3 Mobile	1.679 ms	0 - 148 MB	NPU
ResNeXt50	ONNX	float	Qualcomm® QCS8550 (Proxy)	2.24 ms	0 - 40 MB	NPU
ResNeXt50	ONNX	float	Qualcomm® QCS9075	3.451 ms	0 - 4 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	1.379 ms	0 - 84 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.609 ms	0 - 79 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® 8 Elite Mobile	0.697 ms	0 - 79 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® X2 Elite	0.514 ms	25 - 25 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® X Elite	1.27 ms	25 - 25 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® X Elite	1.27 ms	25 - 25 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.825 ms	0 - 104 MB	NPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS6490	47.935 ms	5 - 23 MB	CPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	1.121 ms	0 - 33 MB	NPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS9075	1.234 ms	0 - 3 MB	NPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCM6690	26.428 ms	2 - 12 MB	CPU
ResNeXt50	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.697 ms	0 - 79 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	20.704 ms	3 - 13 MB	CPU
ResNeXt50	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	20.704 ms	3 - 13 MB	CPU
ResNeXt50	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	1.171 ms	0 - 74 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® 8 Elite Mobile	1.477 ms	0 - 73 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® X2 Elite	1.438 ms	1 - 1 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® X Elite	2.669 ms	1 - 1 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® X Elite	2.669 ms	1 - 1 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	1.794 ms	0 - 136 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	12.035 ms	1 - 71 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	2.563 ms	1 - 162 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA8775P	3.825 ms	1 - 72 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA8775P	3.825 ms	1 - 72 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA8775P	3.825 ms	1 - 72 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS9075	3.637 ms	1 - 3 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	4.031 ms	0 - 113 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA7255P	12.035 ms	1 - 71 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA8295P	4.1 ms	0 - 51 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	1.477 ms	0 - 73 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.539 ms	0 - 75 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 8 Elite Mobile	0.642 ms	0 - 74 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.63 ms	0 - 0 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® X Elite	1.231 ms	0 - 0 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® X Elite	1.231 ms	0 - 0 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.792 ms	0 - 99 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS6490	3.108 ms	0 - 2 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	2.445 ms	0 - 69 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	1.115 ms	0 - 2 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA8775P	1.464 ms	0 - 72 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA8775P	1.464 ms	0 - 72 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA8775P	1.464 ms	0 - 72 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS9075	1.206 ms	0 - 2 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCM6690	7.496 ms	0 - 194 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	1.466 ms	0 - 101 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA7255P	2.445 ms	0 - 69 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA8295P	1.773 ms	0 - 68 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.642 ms	0 - 74 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	1.372 ms	0 - 75 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	1.372 ms	0 - 75 MB	NPU
ResNeXt50	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	1.183 ms	0 - 116 MB	NPU
ResNeXt50	TFLITE	float	Snapdragon® 8 Elite Mobile	1.432 ms	0 - 116 MB	NPU
ResNeXt50	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	1.769 ms	0 - 175 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS8275 (Proxy)	12.055 ms	0 - 112 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS8550 (Proxy)	2.48 ms	0 - 2 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA8775P	3.83 ms	0 - 112 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA8775P	3.83 ms	0 - 112 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA8775P	3.83 ms	0 - 112 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS9075	3.711 ms	0 - 52 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS8450 (Proxy)	3.997 ms	0 - 152 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA7255P	12.055 ms	0 - 112 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA8295P	4.176 ms	0 - 87 MB	NPU
ResNeXt50	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	1.432 ms	0 - 116 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.489 ms	0 - 73 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 8 Elite Mobile	0.563 ms	0 - 73 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.667 ms	0 - 99 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS6490	2.826 ms	0 - 27 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	2.199 ms	0 - 70 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	0.893 ms	0 - 132 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA8775P	1.272 ms	0 - 71 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA8775P	1.272 ms	0 - 71 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA8775P	1.272 ms	0 - 71 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS9075	0.995 ms	0 - 27 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCM6690	7.354 ms	0 - 198 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	1.286 ms	0 - 101 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA7255P	2.199 ms	0 - 70 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA8295P	1.557 ms	0 - 65 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.563 ms	0 - 73 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	1.219 ms	0 - 70 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	1.219 ms	0 - 70 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.