ResNet50: Optimized for Qualcomm Devices

ResNet50 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 ResNet50 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 ResNet50 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 ResNet50 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: 25.5M
• Model size (float): 97.4 MB
• Model size (w8a8): 25.1 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
ResNet50	ONNX	float	Snapdragon® X Elite	2.108 ms	49 - 49 MB	NPU
ResNet50	ONNX	float	Snapdragon® 8 Gen 3 Mobile	1.724 ms	0 - 140 MB	NPU
ResNet50	ONNX	float	Qualcomm® QCS8550 (Proxy)	2.281 ms	0 - 61 MB	NPU
ResNet50	ONNX	float	Qualcomm® QCS9075	3.284 ms	1 - 4 MB	NPU
ResNet50	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	1.387 ms	0 - 111 MB	NPU
ResNet50	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	1.212 ms	0 - 134 MB	NPU
ResNet50	ONNX	w8a8	Snapdragon® X Elite	0.962 ms	25 - 25 MB	NPU
ResNet50	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.864 ms	0 - 147 MB	NPU
ResNet50	ONNX	w8a8	Qualcomm® QCS6490	32.258 ms	9 - 31 MB	CPU
ResNet50	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	1.118 ms	0 - 173 MB	NPU
ResNet50	ONNX	w8a8	Qualcomm® QCS9075	1.205 ms	0 - 3 MB	NPU
ResNet50	ONNX	w8a8	Qualcomm® QCM6690	22.973 ms	1 - 11 MB	CPU
ResNet50	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.716 ms	0 - 110 MB	NPU
ResNet50	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	17.351 ms	6 - 13 MB	CPU
ResNet50	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.663 ms	0 - 114 MB	NPU
ResNet50	QNN_DLC	float	Snapdragon® X Elite	2.379 ms	1 - 1 MB	NPU
ResNet50	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	1.638 ms	0 - 79 MB	NPU
ResNet50	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	10.695 ms	1 - 47 MB	NPU
ResNet50	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	2.233 ms	1 - 3 MB	NPU
ResNet50	QNN_DLC	float	Qualcomm® SA8775P	3.386 ms	1 - 48 MB	NPU
ResNet50	QNN_DLC	float	Qualcomm® QCS9075	3.311 ms	1 - 3 MB	NPU
ResNet50	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	3.739 ms	0 - 62 MB	NPU
ResNet50	QNN_DLC	float	Qualcomm® SA7255P	10.695 ms	1 - 47 MB	NPU
ResNet50	QNN_DLC	float	Qualcomm® SA8295P	3.67 ms	0 - 30 MB	NPU
ResNet50	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	1.325 ms	0 - 48 MB	NPU
ResNet50	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	1.099 ms	1 - 70 MB	NPU
ResNet50	QNN_DLC	w8a8	Snapdragon® X Elite	0.958 ms	0 - 0 MB	NPU
ResNet50	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.687 ms	0 - 70 MB	NPU
ResNet50	QNN_DLC	w8a8	Qualcomm® QCS6490	3.002 ms	0 - 2 MB	NPU
ResNet50	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	1.975 ms	0 - 41 MB	NPU
ResNet50	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	0.902 ms	0 - 183 MB	NPU
ResNet50	QNN_DLC	w8a8	Qualcomm® SA8775P	1.081 ms	0 - 44 MB	NPU
ResNet50	QNN_DLC	w8a8	Qualcomm® QCS9075	0.962 ms	0 - 2 MB	NPU
ResNet50	QNN_DLC	w8a8	Qualcomm® QCM6690	6.501 ms	0 - 162 MB	NPU
ResNet50	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	1.199 ms	0 - 74 MB	NPU
ResNet50	QNN_DLC	w8a8	Qualcomm® SA7255P	1.975 ms	0 - 41 MB	NPU
ResNet50	QNN_DLC	w8a8	Qualcomm® SA8295P	1.445 ms	0 - 39 MB	NPU
ResNet50	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.532 ms	0 - 38 MB	NPU
ResNet50	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	1.259 ms	0 - 47 MB	NPU
ResNet50	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.494 ms	0 - 41 MB	NPU
ResNet50	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	1.6 ms	0 - 121 MB	NPU
ResNet50	TFLITE	float	Qualcomm® QCS8275 (Proxy)	10.624 ms	0 - 82 MB	NPU
ResNet50	TFLITE	float	Qualcomm® QCS8550 (Proxy)	2.198 ms	0 - 2 MB	NPU
ResNet50	TFLITE	float	Qualcomm® SA8775P	14.365 ms	0 - 82 MB	NPU
ResNet50	TFLITE	float	Qualcomm® QCS9075	3.343 ms	0 - 52 MB	NPU
ResNet50	TFLITE	float	Qualcomm® QCS8450 (Proxy)	3.728 ms	0 - 105 MB	NPU
ResNet50	TFLITE	float	Qualcomm® SA7255P	10.624 ms	0 - 82 MB	NPU
ResNet50	TFLITE	float	Qualcomm® SA8295P	3.686 ms	0 - 67 MB	NPU
ResNet50	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	1.302 ms	0 - 78 MB	NPU
ResNet50	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	1.1 ms	0 - 103 MB	NPU
ResNet50	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.576 ms	0 - 72 MB	NPU
ResNet50	TFLITE	w8a8	Qualcomm® QCS6490	2.612 ms	0 - 27 MB	NPU
ResNet50	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	1.733 ms	0 - 40 MB	NPU
ResNet50	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	0.767 ms	0 - 31 MB	NPU
ResNet50	TFLITE	w8a8	Qualcomm® SA8775P	0.96 ms	0 - 43 MB	NPU
ResNet50	TFLITE	w8a8	Qualcomm® QCS9075	0.813 ms	0 - 27 MB	NPU
ResNet50	TFLITE	w8a8	Qualcomm® QCM6690	6.25 ms	0 - 159 MB	NPU
ResNet50	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	1.059 ms	0 - 74 MB	NPU
ResNet50	TFLITE	w8a8	Qualcomm® SA7255P	1.733 ms	0 - 40 MB	NPU
ResNet50	TFLITE	w8a8	Qualcomm® SA8295P	1.314 ms	0 - 37 MB	NPU
ResNet50	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.491 ms	0 - 38 MB	NPU
ResNet50	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	1.088 ms	0 - 46 MB	NPU
ResNet50	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.445 ms	0 - 41 MB	NPU

License

• The license for the original implementation of ResNet50 can be found here.

References

• Deep Residual Learning for Image Recognition
• 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.