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 1.24.1	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 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® X2 Elite	0.513 ms	23 - 23 MB	NPU
ResNet18	ONNX	float	Snapdragon® X Elite	1.179 ms	22 - 22 MB	NPU
ResNet18	ONNX	float	Snapdragon® 8 Gen 3 Mobile	0.778 ms	0 - 35 MB	NPU
ResNet18	ONNX	float	Qualcomm® QCS8550 (Proxy)	1.066 ms	0 - 40 MB	NPU
ResNet18	ONNX	float	Qualcomm® QCS9075	1.806 ms	1 - 3 MB	NPU
ResNet18	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	0.63 ms	0 - 20 MB	NPU
ResNet18	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	0.556 ms	1 - 26 MB	NPU
ResNet18	ONNX	w8a8	Snapdragon® X2 Elite	0.266 ms	11 - 11 MB	NPU
ResNet18	ONNX	w8a8	Snapdragon® X Elite	0.627 ms	11 - 11 MB	NPU
ResNet18	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.415 ms	0 - 48 MB	NPU
ResNet18	ONNX	w8a8	Qualcomm® QCS6490	13.609 ms	7 - 23 MB	CPU
ResNet18	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	0.663 ms	0 - 32 MB	NPU
ResNet18	ONNX	w8a8	Qualcomm® QCS9075	0.669 ms	0 - 3 MB	NPU
ResNet18	ONNX	w8a8	Qualcomm® QCM6690	11.322 ms	6 - 13 MB	CPU
ResNet18	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.367 ms	0 - 27 MB	NPU
ResNet18	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	8.668 ms	8 - 15 MB	CPU
ResNet18	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.342 ms	0 - 26 MB	NPU
ResNet18	QNN_DLC	float	Snapdragon® X2 Elite	0.694 ms	1 - 1 MB	NPU
ResNet18	QNN_DLC	float	Snapdragon® X Elite	1.468 ms	1 - 1 MB	NPU
ResNet18	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	0.911 ms	0 - 35 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	5.782 ms	1 - 24 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	1.328 ms	1 - 37 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® SA8775P	1.947 ms	1 - 27 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® QCS9075	2.064 ms	1 - 3 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	2.431 ms	0 - 37 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® SA7255P	5.782 ms	1 - 24 MB	NPU
ResNet18	QNN_DLC	float	Qualcomm® SA8295P	2.308 ms	0 - 19 MB	NPU
ResNet18	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	0.7 ms	0 - 23 MB	NPU
ResNet18	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	0.586 ms	1 - 28 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.311 ms	0 - 0 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® X Elite	0.617 ms	0 - 0 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.391 ms	0 - 47 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS6490	1.802 ms	2 - 4 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	1.269 ms	0 - 23 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	0.524 ms	0 - 2 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® SA8775P	2.697 ms	0 - 23 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS9075	0.642 ms	2 - 4 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCM6690	3.234 ms	0 - 31 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	0.745 ms	0 - 49 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® SA7255P	1.269 ms	0 - 23 MB	NPU
ResNet18	QNN_DLC	w8a8	Qualcomm® SA8295P	0.907 ms	0 - 21 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.304 ms	0 - 26 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	0.682 ms	0 - 31 MB	NPU
ResNet18	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.264 ms	0 - 26 MB	NPU
ResNet18	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	0.92 ms	0 - 64 MB	NPU
ResNet18	TFLITE	float	Qualcomm® QCS8275 (Proxy)	5.745 ms	0 - 28 MB	NPU
ResNet18	TFLITE	float	Qualcomm® QCS8550 (Proxy)	1.342 ms	0 - 2 MB	NPU
ResNet18	TFLITE	float	Qualcomm® SA8775P	1.944 ms	0 - 30 MB	NPU
ResNet18	TFLITE	float	Qualcomm® QCS9075	2.033 ms	0 - 25 MB	NPU
ResNet18	TFLITE	float	Qualcomm® QCS8450 (Proxy)	2.467 ms	0 - 61 MB	NPU
ResNet18	TFLITE	float	Qualcomm® SA7255P	5.745 ms	0 - 28 MB	NPU
ResNet18	TFLITE	float	Qualcomm® SA8295P	2.3 ms	0 - 22 MB	NPU
ResNet18	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	0.705 ms	0 - 26 MB	NPU
ResNet18	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	0.589 ms	0 - 30 MB	NPU
ResNet18	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.283 ms	0 - 46 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS6490	1.527 ms	0 - 13 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	1.019 ms	0 - 23 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	0.385 ms	0 - 2 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® SA8775P	0.573 ms	0 - 25 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS9075	0.508 ms	0 - 13 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCM6690	2.781 ms	0 - 31 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	0.599 ms	0 - 48 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® SA7255P	1.019 ms	0 - 23 MB	NPU
ResNet18	TFLITE	w8a8	Qualcomm® SA8295P	0.747 ms	0 - 21 MB	NPU
ResNet18	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.241 ms	0 - 25 MB	NPU
ResNet18	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	0.548 ms	0 - 30 MB	NPU
ResNet18	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.22 ms	0 - 26 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.