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 1.24.1	Download
ONNX	w8a8	Universal	QAIRT 2.42, ONNX Runtime 1.24.1	Download
QNN_DLC	float	Universal	QAIRT 2.43	Download
QNN_DLC	w8a8	Universal	QAIRT 2.43	Download
TFLITE	float	Universal	QAIRT 2.43, TFLite 2.17.0	Download
TFLITE	w8a8	Universal	QAIRT 2.43, TFLite 2.17.0	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® X2 Elite	1.091 ms	50 - 50 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® X Elite	2.428 ms	49 - 49 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® 8 Gen 3 Mobile	1.684 ms	0 - 147 MB	NPU
ResNeXt50	ONNX	float	Qualcomm® QCS8550 (Proxy)	2.26 ms	1 - 8 MB	NPU
ResNeXt50	ONNX	float	Qualcomm® QCS9075	3.448 ms	1 - 4 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	1.378 ms	0 - 83 MB	NPU
ResNeXt50	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	1.145 ms	1 - 85 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® X2 Elite	0.51 ms	25 - 25 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® X Elite	1.277 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	46.205 ms	5 - 20 MB	CPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	1.103 ms	0 - 32 MB	NPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCS9075	1.239 ms	0 - 3 MB	NPU
ResNeXt50	ONNX	w8a8	Qualcomm® QCM6690	26.367 ms	3 - 12 MB	CPU
ResNeXt50	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.692 ms	0 - 78 MB	NPU
ResNeXt50	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	20.796 ms	4 - 12 MB	CPU
ResNeXt50	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.604 ms	0 - 78 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® X2 Elite	1.471 ms	1 - 1 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® X Elite	2.689 ms	1 - 1 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	1.784 ms	0 - 138 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	11.891 ms	1 - 76 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	2.532 ms	1 - 66 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA8775P	3.825 ms	1 - 76 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS9075	3.658 ms	3 - 5 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	4.02 ms	0 - 115 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA7255P	11.891 ms	1 - 76 MB	NPU
ResNeXt50	QNN_DLC	float	Qualcomm® SA8295P	4.12 ms	0 - 53 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	1.454 ms	0 - 75 MB	NPU
ResNeXt50	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	1.183 ms	1 - 77 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.641 ms	0 - 0 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® X Elite	1.233 ms	0 - 0 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.8 ms	0 - 99 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS6490	3.077 ms	0 - 2 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	2.476 ms	0 - 69 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	1.09 ms	0 - 2 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA8775P	1.491 ms	0 - 70 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS9075	1.236 ms	0 - 2 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCM6690	7.616 ms	0 - 193 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	1.523 ms	0 - 99 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA7255P	2.476 ms	0 - 69 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Qualcomm® SA8295P	1.774 ms	0 - 66 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.621 ms	0 - 71 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	1.386 ms	0 - 73 MB	NPU
ResNeXt50	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.544 ms	0 - 72 MB	NPU
ResNeXt50	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	1.757 ms	0 - 181 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS8275 (Proxy)	11.892 ms	0 - 124 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS8550 (Proxy)	2.494 ms	0 - 62 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA8775P	3.818 ms	0 - 124 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS9075	3.727 ms	0 - 52 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® QCS8450 (Proxy)	4.046 ms	0 - 162 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA7255P	11.892 ms	0 - 124 MB	NPU
ResNeXt50	TFLITE	float	Qualcomm® SA8295P	4.156 ms	0 - 104 MB	NPU
ResNeXt50	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	1.426 ms	0 - 128 MB	NPU
ResNeXt50	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	1.197 ms	0 - 126 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.655 ms	0 - 99 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS6490	2.868 ms	0 - 27 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	2.226 ms	0 - 67 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	0.919 ms	0 - 51 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA8775P	1.294 ms	0 - 68 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS9075	1.02 ms	0 - 27 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCM6690	7.33 ms	0 - 196 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	1.312 ms	0 - 97 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA7255P	2.226 ms	0 - 67 MB	NPU
ResNeXt50	TFLITE	w8a8	Qualcomm® SA8295P	1.567 ms	0 - 62 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.562 ms	0 - 70 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	1.232 ms	0 - 67 MB	NPU
ResNeXt50	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.483 ms	0 - 71 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.