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 1.25.0	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 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	3.055 ms	1 - 197 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® X2 Elite	3.109 ms	211 - 211 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® X Elite	6.459 ms	172 - 172 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® 8 Gen 3 Mobile	4.493 ms	0 - 369 MB	NPU
ResNeXt101	ONNX	float	Qualcomm® QCS8550 (Proxy)	6.681 ms	0 - 196 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	3.892 ms	0 - 192 MB	NPU
ResNeXt101	ONNX	float	Qualcomm® QCS9075	9.737 ms	1 - 46 MB	NPU
ResNeXt101	ONNX	float	Qualcomm® QCS8750	3.892 ms	0 - 192 MB	NPU
ResNeXt101	ONNX	float	Qualcomm® QCS7181	6.459 ms	172 - 172 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.566 ms	0 - 222 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® X2 Elite	1.401 ms	212 - 212 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® X Elite	2.904 ms	149 - 149 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	2.132 ms	0 - 264 MB	NPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS6490	113.635 ms	10 - 53 MB	CPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	2.889 ms	0 - 147 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	68.492 ms	0 - 13 MB	CPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCM6690	73.79 ms	14 - 30 MB	CPU
ResNeXt101	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.903 ms	0 - 217 MB	NPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS9075	3.124 ms	0 - 45 MB	NPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS7790	68.492 ms	0 - 13 MB	CPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS8750	1.903 ms	0 - 217 MB	NPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS7181	2.904 ms	149 - 149 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	3.128 ms	0 - 188 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® X2 Elite	3.604 ms	1 - 1 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® X Elite	6.917 ms	1 - 1 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	4.664 ms	0 - 367 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS8275	35.986 ms	1 - 181 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	6.811 ms	1 - 3 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA8775P	10.287 ms	1 - 194 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA8650P	10.287 ms	1 - 194 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA8255P	10.287 ms	1 - 194 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	9.953 ms	0 - 311 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA8295P	10.944 ms	1 - 135 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	3.977 ms	1 - 182 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA7255P	35.986 ms	1 - 181 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS9075	9.973 ms	3 - 5 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS8750	3.977 ms	1 - 182 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS7181	6.917 ms	1 - 1 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.501 ms	0 - 216 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® X2 Elite	1.693 ms	0 - 0 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® X Elite	3.111 ms	0 - 0 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	2.113 ms	0 - 258 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS6490	9.412 ms	2 - 4 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS8275	6.583 ms	0 - 214 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	2.944 ms	0 - 2 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA8775P	3.496 ms	0 - 215 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA8650P	3.496 ms	0 - 215 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA8255P	3.496 ms	0 - 215 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA7255P	6.583 ms	0 - 214 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	4.069 ms	0 - 244 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA8295P	4.302 ms	0 - 217 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCM6690	31.66 ms	0 - 372 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.821 ms	0 - 211 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS9075	3.111 ms	0 - 2 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	4.001 ms	0 - 259 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS7790	4.069 ms	0 - 244 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS8750	1.821 ms	0 - 211 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS7181	3.111 ms	0 - 0 MB	NPU
ResNeXt101	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	3.112 ms	0 - 364 MB	NPU
ResNeXt101	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	4.619 ms	0 - 531 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS8275	35.977 ms	1 - 359 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS8550 (Proxy)	6.772 ms	0 - 3 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA8775P	10.252 ms	0 - 358 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA8650P	10.252 ms	0 - 358 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA8255P	10.252 ms	0 - 358 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS8450 (Proxy)	9.875 ms	0 - 478 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA8295P	10.989 ms	0 - 305 MB	NPU
ResNeXt101	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	3.953 ms	0 - 355 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA7255P	35.977 ms	1 - 359 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS9075	10.028 ms	0 - 174 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS8750	3.953 ms	0 - 355 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.465 ms	0 - 213 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	1.977 ms	0 - 248 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS6490	8.94 ms	0 - 89 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS8275	6.324 ms	0 - 213 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	2.744 ms	0 - 3 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA8775P	3.314 ms	0 - 214 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA8650P	3.314 ms	0 - 214 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA8255P	3.314 ms	0 - 214 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA7255P	6.324 ms	0 - 213 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	3.966 ms	0 - 243 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA8295P	4.089 ms	0 - 216 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCM6690	30.386 ms	0 - 374 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.728 ms	0 - 208 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS9075	2.921 ms	0 - 89 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	3.824 ms	0 - 255 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS7790	3.966 ms	0 - 243 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS8750	1.728 ms	0 - 208 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.