GoogLeNet: Optimized for Qualcomm Devices

GoogLeNet 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 GoogLeNet 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 GoogLeNet 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 GoogLeNet 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: 6.62M
• Model size (float): 25.3 MB
• Model size (w8a8): 6.54 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
GoogLeNet	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	0.398 ms	0 - 31 MB	NPU
GoogLeNet	ONNX	float	Snapdragon® X2 Elite	0.432 ms	13 - 13 MB	NPU
GoogLeNet	ONNX	float	Snapdragon® X Elite	1.001 ms	13 - 13 MB	NPU
GoogLeNet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	0.564 ms	0 - 46 MB	NPU
GoogLeNet	ONNX	float	Qualcomm® QCS8550 (Proxy)	0.842 ms	0 - 2 MB	NPU
GoogLeNet	ONNX	float	Qualcomm® QCS9075	1.543 ms	1 - 3 MB	NPU
GoogLeNet	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	0.474 ms	0 - 31 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.182 ms	0 - 35 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® X2 Elite	0.17 ms	7 - 7 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® X Elite	0.41 ms	7 - 7 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.226 ms	0 - 47 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® QCS6490	13.577 ms	6 - 17 MB	CPU
GoogLeNet	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	0.299 ms	0 - 113 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® QCS9075	0.431 ms	0 - 3 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® QCM6690	9.366 ms	7 - 15 MB	CPU
GoogLeNet	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.195 ms	0 - 35 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	7.141 ms	6 - 15 MB	CPU
GoogLeNet	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	0.379 ms	1 - 32 MB	NPU
GoogLeNet	QNN_DLC	float	Snapdragon® X2 Elite	0.528 ms	1 - 1 MB	NPU
GoogLeNet	QNN_DLC	float	Snapdragon® X Elite	0.98 ms	1 - 1 MB	NPU
GoogLeNet	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	0.554 ms	0 - 44 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	4.969 ms	1 - 28 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	0.805 ms	1 - 2 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® SA8775P	1.526 ms	1 - 30 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® QCS9075	1.535 ms	3 - 5 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	1.757 ms	0 - 45 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® SA7255P	4.969 ms	1 - 28 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® SA8295P	1.788 ms	1 - 26 MB	NPU
GoogLeNet	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	0.451 ms	0 - 27 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.137 ms	0 - 31 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.238 ms	0 - 0 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® X Elite	0.358 ms	0 - 0 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.188 ms	0 - 42 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® QCS6490	1.055 ms	0 - 2 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	0.806 ms	0 - 28 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	0.236 ms	0 - 2 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® SA8775P	0.416 ms	0 - 29 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® QCS9075	0.331 ms	0 - 2 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® QCM6690	2.248 ms	0 - 28 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	0.411 ms	0 - 44 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® SA7255P	0.806 ms	0 - 28 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® SA8295P	0.643 ms	0 - 25 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.151 ms	0 - 31 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	0.334 ms	0 - 29 MB	NPU
GoogLeNet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	0.383 ms	0 - 37 MB	NPU
GoogLeNet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	0.549 ms	0 - 55 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® QCS8275 (Proxy)	5.005 ms	0 - 34 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® QCS8550 (Proxy)	0.811 ms	0 - 1 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® SA8775P	1.538 ms	0 - 37 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® QCS9075	1.529 ms	0 - 16 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® QCS8450 (Proxy)	1.771 ms	0 - 57 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® SA7255P	5.005 ms	0 - 34 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® SA8295P	1.776 ms	0 - 32 MB	NPU
GoogLeNet	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	0.45 ms	0 - 34 MB	NPU
GoogLeNet	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.144 ms	0 - 30 MB	NPU
GoogLeNet	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.179 ms	0 - 42 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® QCS6490	0.891 ms	0 - 8 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	0.849 ms	0 - 27 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	0.246 ms	0 - 1 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® SA8775P	1.527 ms	0 - 27 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® QCS9075	0.354 ms	0 - 9 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® QCM6690	2.236 ms	0 - 28 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	0.419 ms	0 - 42 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® SA7255P	0.849 ms	0 - 27 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® SA8295P	0.645 ms	0 - 25 MB	NPU
GoogLeNet	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	0.153 ms	0 - 26 MB	NPU
GoogLeNet	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	0.329 ms	0 - 28 MB	NPU

License

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

References

• Going Deeper with Convolutions
• 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.