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 2.37, ONNX Runtime 1.23.0	Download
ONNX	w8a8	Universal	QAIRT 2.37, ONNX Runtime 1.23.0	Download
QNN_DLC	float	Universal	QAIRT 2.42	Download
QNN_DLC	w8a8	Universal	QAIRT 2.42	Download
TFLITE	float	Universal	QAIRT 2.42, TFLite 2.17.0	Download
TFLITE	w8a8	Universal	QAIRT 2.42, 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® X Elite	6.573 ms	172 - 172 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® 8 Gen 3 Mobile	4.726 ms	0 - 405 MB	NPU
ResNeXt101	ONNX	float	Qualcomm® QCS8550 (Proxy)	6.715 ms	0 - 194 MB	NPU
ResNeXt101	ONNX	float	Qualcomm® QCS9075	10.044 ms	0 - 4 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	4.052 ms	0 - 226 MB	NPU
ResNeXt101	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	3.262 ms	0 - 233 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® X Elite	3.135 ms	87 - 87 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	2.372 ms	0 - 330 MB	NPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS6490	112.485 ms	3 - 52 MB	CPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	3.18 ms	0 - 98 MB	NPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCS9075	3.411 ms	0 - 3 MB	NPU
ResNeXt101	ONNX	w8a8	Qualcomm® QCM6690	72.672 ms	0 - 12 MB	CPU
ResNeXt101	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	2.153 ms	0 - 285 MB	NPU
ResNeXt101	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	67.614 ms	0 - 12 MB	CPU
ResNeXt101	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.741 ms	0 - 288 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® X Elite	6.785 ms	1 - 1 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	4.852 ms	0 - 384 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	36.113 ms	1 - 203 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	6.579 ms	1 - 5 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA8775P	10.261 ms	1 - 190 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS9075	9.977 ms	3 - 5 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	9.874 ms	0 - 312 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA7255P	36.113 ms	1 - 203 MB	NPU
ResNeXt101	QNN_DLC	float	Qualcomm® SA8295P	10.903 ms	1 - 140 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	3.984 ms	1 - 189 MB	NPU
ResNeXt101	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	3.111 ms	0 - 193 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® X Elite	3.062 ms	0 - 0 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	2.125 ms	0 - 252 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS6490	9.111 ms	2 - 4 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	6.538 ms	0 - 206 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	2.85 ms	0 - 2 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA8775P	3.469 ms	0 - 208 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS9075	3.118 ms	0 - 2 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCM6690	26.699 ms	0 - 368 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	3.877 ms	0 - 251 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA7255P	6.538 ms	0 - 206 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Qualcomm® SA8295P	4.239 ms	0 - 210 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.808 ms	0 - 204 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	4.088 ms	0 - 245 MB	NPU
ResNeXt101	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.497 ms	0 - 210 MB	NPU
ResNeXt101	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	4.673 ms	0 - 562 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS8275 (Proxy)	35.918 ms	0 - 397 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS8550 (Proxy)	6.91 ms	0 - 3 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA8775P	10.374 ms	0 - 408 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS9075	10.026 ms	0 - 174 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® QCS8450 (Proxy)	9.84 ms	0 - 517 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA7255P	35.918 ms	0 - 397 MB	NPU
ResNeXt101	TFLITE	float	Qualcomm® SA8295P	10.865 ms	0 - 342 MB	NPU
ResNeXt101	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	3.917 ms	0 - 385 MB	NPU
ResNeXt101	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	3.105 ms	0 - 395 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	1.971 ms	0 - 242 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS6490	8.997 ms	0 - 88 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	6.265 ms	0 - 218 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	2.647 ms	0 - 3 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA8775P	3.285 ms	0 - 219 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS9075	2.926 ms	0 - 89 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCM6690	24.894 ms	0 - 363 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	3.873 ms	0 - 253 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA7255P	6.265 ms	0 - 218 MB	NPU
ResNeXt101	TFLITE	w8a8	Qualcomm® SA8295P	4.065 ms	0 - 219 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.774 ms	0 - 203 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	3.981 ms	0 - 294 MB	NPU
ResNeXt101	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.447 ms	0 - 213 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.