ResNet18: Optimized for Qualcomm Devices

ResNet18 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 ResNet18 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 ResNet18 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 ResNet18 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: 11.7M
• Model size (float): 44.6 MB
• Model size (w8a8): 11.3 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
ResNet18	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	0.552 ms	0 - 26 MB	NPU
ResNet18	ONNX	float	Snapdragon® X2 Elite	0.513 ms	211 - 211 MB	NPU
ResNet18	ONNX	float	Snapdragon® X Elite	1.048 ms	149 - 149 MB	NPU
ResNet18	ONNX	float	Snapdragon® 8 Gen 3 Mobile	0.774 ms	0 - 41 MB	NPU
ResNet18	ONNX	float	Qualcomm® QCS8550 (Proxy)	1.072 ms	0 - 151 MB	NPU
ResNet18	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	0.629 ms	0 - 21 MB	NPU
ResNet18	ONNX	float	Qualcomm® QCS9075	1.802 ms	1 - 46 MB	NPU
ResNet18	ONNX	float	Qualcomm® QCS8750	0.629 ms	0 - 21 MB	NPU
ResNet18	ONNX	float	Qualcomm® QCS7181	1.048 ms	149 - 149 MB	NPU
ResNet18	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.337 ms	0 - 29 MB	NPU
ResNet18	ONNX	w8a8	Snapdragon® X2 Elite	0.267 ms	212 - 212 MB	NPU
ResNet18	ONNX	w8a8	Snapdragon® X Elite	0.505 ms	149 - 149 MB	NPU
ResNet18	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.418 ms	0 - 54 MB	NPU
ResNet18	ONNX	w8a8	Qualcomm® QCS6490	13.521 ms	6 - 20 MB	CPU
ResNet18	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	0.558 ms	0 - 27 MB	NPU
ResNet18	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	8.674 ms	8 - 15 MB	CPU
ResNet18	ONNX	w8a8	Qualcomm® QCM6690	11.406 ms	6 - 12 MB	CPU
ResNet18	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.364 ms	0 - 29 MB	NPU
ResNet18	ONNX	w8a8	Qualcomm® QCS9075	0.649 ms	0 - 45 MB	NPU
ResNet18	ONNX	w8a8	Qualcomm® QCS7790	8.674 ms	8 - 15 MB	CPU
ResNet18	ONNX	w8a8	Qualcomm® QCS8750	0.364 ms	0 - 29 MB	NPU
ResNet18	ONNX	w8a8	Qualcomm® QCS7181	0.505 ms	149 - 149 MB	NPU
ResNet18	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	0.59 ms	1 - 27 MB	NPU
ResNet18	QNN_DLC	float	Snapdragon® X2 Elite	0.665 ms	1 - 1 MB	NPU
ResNet18	QNN_DLC	float	Snapdragon® X Elite	1.47 ms	1 - 1 MB	NPU
ResNet18	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	0.907 ms	0 - 35 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® QCS8275	5.86 ms	1 - 24 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	1.325 ms	1 - 50 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® SA8775P	1.927 ms	1 - 26 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® SA8650P	1.927 ms	1 - 26 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® SA8255P	1.927 ms	1 - 26 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	2.434 ms	0 - 38 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® SA8295P	2.305 ms	0 - 20 MB	NPU
ResNet18	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	0.718 ms	0 - 27 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® SA7255P	5.86 ms	1 - 24 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® QCS9075	2.065 ms	1 - 3 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® QCS8750	0.718 ms	0 - 27 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® QCS7181	1.47 ms	1 - 1 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.269 ms	0 - 27 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.331 ms	0 - 0 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® X Elite	0.603 ms	0 - 0 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.411 ms	0 - 47 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS6490	1.816 ms	2 - 4 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS8275	1.321 ms	0 - 24 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	0.545 ms	0 - 2 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® SA8775P	0.712 ms	0 - 26 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® SA8650P	0.712 ms	0 - 26 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® SA8255P	0.712 ms	0 - 26 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® SA7255P	1.321 ms	0 - 24 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	0.68 ms	0 - 32 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® SA8295P	0.969 ms	0 - 22 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCM6690	3.233 ms	0 - 33 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.31 ms	0 - 27 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS9075	0.617 ms	2 - 4 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	0.758 ms	0 - 49 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS7790	0.68 ms	0 - 32 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS8750	0.31 ms	0 - 27 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS7181	0.603 ms	0 - 0 MB	NPU
ResNet18	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	0.587 ms	0 - 30 MB	NPU
ResNet18	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	0.916 ms	0 - 55 MB	NPU
ResNet18	TFLITE	float	Qualcomm® QCS8275	5.842 ms	0 - 26 MB	NPU
ResNet18	TFLITE	float	Qualcomm® QCS8550 (Proxy)	1.32 ms	0 - 2 MB	NPU
ResNet18	TFLITE	float	Qualcomm® SA8775P	1.916 ms	0 - 29 MB	NPU
ResNet18	TFLITE	float	Qualcomm® SA8650P	1.916 ms	0 - 29 MB	NPU
ResNet18	TFLITE	float	Qualcomm® SA8255P	1.916 ms	0 - 29 MB	NPU
ResNet18	TFLITE	float	Qualcomm® QCS8450 (Proxy)	2.458 ms	0 - 59 MB	NPU
ResNet18	TFLITE	float	Qualcomm® SA8295P	2.279 ms	0 - 22 MB	NPU
ResNet18	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	0.704 ms	0 - 31 MB	NPU
ResNet18	TFLITE	float	Qualcomm® SA7255P	5.842 ms	0 - 26 MB	NPU
ResNet18	TFLITE	float	Qualcomm® QCS9075	2.04 ms	0 - 25 MB	NPU
ResNet18	TFLITE	float	Qualcomm® QCS8750	0.704 ms	0 - 31 MB	NPU
ResNet18	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.228 ms	0 - 26 MB	NPU
ResNet18	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.293 ms	0 - 47 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS6490	1.497 ms	0 - 13 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS8275	1.025 ms	0 - 23 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	0.386 ms	0 - 72 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® SA8775P	0.558 ms	0 - 25 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® SA8650P	0.558 ms	0 - 25 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® SA8255P	0.558 ms	0 - 25 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® SA7255P	1.025 ms	0 - 23 MB	NPU
ResNet18	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	0.551 ms	0 - 31 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® SA8295P	0.763 ms	0 - 22 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCM6690	2.797 ms	0 - 31 MB	NPU
ResNet18	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.246 ms	0 - 22 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS9075	0.466 ms	0 - 13 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	0.608 ms	0 - 48 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS7790	0.551 ms	0 - 31 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS8750	0.246 ms	0 - 22 MB	NPU

License

• The license for the original implementation of ResNet18 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.