ResNeXt101: Optimized for Qualcomm Devices

ResNeXt101 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 ResNeXt101 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 ResNeXt101 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 ResNeXt101 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: 88.7M
• Model size (float): 338 MB
• Model size (w8a8): 87.3 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
ResNeXt101	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	3.05 ms	1 - 203 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® X2 Elite	3.101 ms	173 - 173 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® X Elite	6.692 ms	172 - 172 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® 8 Gen 3 Mobile	4.503 ms	0 - 373 MB	NPU
ResNeXt101	ONNX	float	Qualcomm® QCS8550 (Proxy)	6.643 ms	0 - 199 MB	NPU
ResNeXt101	ONNX	float	Qualcomm® QCS9075	9.779 ms	0 - 4 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	3.88 ms	0 - 175 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.576 ms	0 - 215 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® X2 Elite	1.386 ms	87 - 87 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® X Elite	3.139 ms	87 - 87 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	2.154 ms	0 - 259 MB	NPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS6490	112.681 ms	3 - 36 MB	CPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	2.937 ms	0 - 100 MB	NPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS9075	3.135 ms	0 - 3 MB	NPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCM6690	73.093 ms	0 - 12 MB	CPU
ResNeXt101	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.859 ms	0 - 224 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	68.135 ms	0 - 12 MB	CPU
ResNeXt101	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	3.093 ms	1 - 195 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® X2 Elite	3.768 ms	1 - 1 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® X Elite	6.939 ms	1 - 1 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	4.642 ms	0 - 371 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	35.894 ms	1 - 191 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	6.836 ms	1 - 3 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA8775P	10.31 ms	1 - 203 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS9075	9.988 ms	1 - 3 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	9.882 ms	0 - 313 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA7255P	35.894 ms	1 - 191 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA8295P	10.937 ms	1 - 140 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	3.993 ms	1 - 189 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.525 ms	0 - 210 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® X2 Elite	1.765 ms	0 - 0 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® X Elite	3.078 ms	0 - 0 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	2.138 ms	0 - 250 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS6490	9.267 ms	0 - 2 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	6.517 ms	0 - 206 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	2.847 ms	0 - 4 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA8775P	3.507 ms	0 - 207 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS9075	3.137 ms	0 - 2 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCM6690	33.668 ms	0 - 370 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	3.902 ms	0 - 250 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA7255P	6.517 ms	0 - 206 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA8295P	4.257 ms	0 - 209 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.809 ms	0 - 207 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	4.065 ms	0 - 241 MB	NPU
ResNeXt101	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	3.093 ms	0 - 401 MB	NPU
ResNeXt101	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	4.642 ms	0 - 580 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS8275 (Proxy)	35.913 ms	0 - 396 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS8550 (Proxy)	6.793 ms	0 - 2 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA8775P	45.167 ms	0 - 396 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS9075	10.041 ms	0 - 174 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS8450 (Proxy)	9.882 ms	0 - 511 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA7255P	35.913 ms	0 - 396 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA8295P	10.968 ms	0 - 336 MB	NPU
ResNeXt101	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	3.96 ms	0 - 398 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.472 ms	0 - 211 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	2.001 ms	0 - 257 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS6490	8.87 ms	0 - 88 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	6.29 ms	0 - 210 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	2.664 ms	0 - 4 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA8775P	3.349 ms	0 - 211 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS9075	2.959 ms	0 - 89 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCM6690	28.419 ms	0 - 370 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	3.719 ms	0 - 249 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA7255P	6.29 ms	0 - 210 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA8295P	4.076 ms	0 - 213 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.773 ms	0 - 209 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	3.96 ms	0 - 241 MB	NPU

License

• The license for the original implementation of ResNeXt101 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.