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 2.45, ONNX Runtime 1.27.1	Download
ONNX	w8a8	Universal	QAIRT 2.45, ONNX Runtime 1.27.1	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 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:

• Input resolution: 224x224
• Model checkpoint: Imagenet
• Model size (float): 25.3 MB
• Model size (w8a8): 6.54 MB
• Number of parameters: 6.62M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
GoogLeNet	ONNX	float	Snapdragon® X2 Elite	0.431 ms	2 - 2 MB	NPU
GoogLeNet	ONNX	float	Snapdragon® X Elite	0.869 ms	13 - 13 MB	NPU
GoogLeNet	ONNX	float	Snapdragon® 8 Gen 3 Mobile	0.561 ms	0 - 46 MB	NPU
GoogLeNet	ONNX	float	Snapdragon® 8 Gen 1 Mobile	1.581 ms	1 - 51 MB	NPU
GoogLeNet	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	1.533 ms	1 - 4 MB	NPU
GoogLeNet	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.842 ms	0 - 81 MB	NPU
GoogLeNet	ONNX	float	Qualcomm® QCS8450	1.581 ms	1 - 51 MB	NPU
GoogLeNet	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	1.494 ms	1 - 3 MB	NPU
GoogLeNet	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	0.869 ms	13 - 13 MB	NPU
GoogLeNet	ONNX	float	Qualcomm® Dragonwing™ Q-8750	0.467 ms	0 - 27 MB	NPU
GoogLeNet	ONNX	float	Snapdragon® 8 Elite Mobile	0.467 ms	0 - 27 MB	NPU
GoogLeNet	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	0.393 ms	0 - 27 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® X2 Elite	0.161 ms	1 - 1 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® X Elite	0.293 ms	7 - 7 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.217 ms	0 - 56 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® 8 Gen 1 Mobile	0.492 ms	0 - 53 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® Dragonwing™ QCS6490	1.185 ms	0 - 3 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-8275	0.382 ms	0 - 4 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.307 ms	0 - 9 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® QCS8450	0.492 ms	0 - 53 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-9075	0.428 ms	0 - 3 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-X7181	0.293 ms	7 - 7 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® Dragonwing™ Q-6690	2.314 ms	0 - 36 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® Dragonwing™ Q-7790	0.35 ms	0 - 36 MB	NPU
GoogLeNet	ONNX	w8a8	Qualcomm® Dragonwing™ Q-8750	0.198 ms	0 - 38 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® 8 Elite Mobile	0.198 ms	0 - 38 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.188 ms	0 - 35 MB	NPU
GoogLeNet	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	0.35 ms	0 - 36 MB	NPU
GoogLeNet	QNN_DLC	float	Snapdragon® X2 Elite	0.501 ms	1 - 1 MB	NPU
GoogLeNet	QNN_DLC	float	Snapdragon® X Elite	0.959 ms	1 - 1 MB	NPU
GoogLeNet	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	0.565 ms	0 - 43 MB	NPU
GoogLeNet	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	1.75 ms	0 - 49 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	1.464 ms	1 - 4 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.813 ms	1 - 2 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® SA8775P	1.528 ms	1 - 29 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® SA8650P	1.528 ms	1 - 29 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® SA8255P	1.528 ms	1 - 29 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® QCS8450	1.75 ms	0 - 49 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	1.417 ms	1 - 3 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	0.959 ms	1 - 1 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	0.454 ms	0 - 28 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® SA7255P	4.992 ms	1 - 27 MB	NPU
GoogLeNet	QNN_DLC	float	Qualcomm® SA8295P	1.787 ms	1 - 25 MB	NPU
GoogLeNet	QNN_DLC	float	Snapdragon® 8 Elite Mobile	0.454 ms	0 - 28 MB	NPU
GoogLeNet	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	0.376 ms	1 - 33 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.221 ms	0 - 0 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® X Elite	0.346 ms	1 - 1 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.19 ms	0 - 45 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® 8 Gen 1 Mobile	0.416 ms	0 - 49 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS6490	1.174 ms	2 - 4 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-8275	0.343 ms	0 - 3 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.246 ms	0 - 20 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® SA8775P	0.408 ms	0 - 31 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® SA8650P	0.408 ms	0 - 31 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® SA8255P	0.408 ms	0 - 31 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® QCS8450	0.416 ms	0 - 49 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-9075	0.328 ms	0 - 2 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-X7181	0.346 ms	1 - 1 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-6690	2.355 ms	0 - 32 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-7790	0.33 ms	0 - 31 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-8750	0.156 ms	0 - 29 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® SA7255P	0.804 ms	0 - 29 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Qualcomm® SA8295P	0.627 ms	0 - 27 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® 8 Elite Mobile	0.156 ms	0 - 29 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.139 ms	0 - 29 MB	NPU
GoogLeNet	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	0.33 ms	0 - 31 MB	NPU
GoogLeNet	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	0.562 ms	0 - 54 MB	NPU
GoogLeNet	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	1.756 ms	0 - 60 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	1.482 ms	0 - 16 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.821 ms	0 - 2 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® SA8775P	1.562 ms	0 - 35 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® SA8650P	1.562 ms	0 - 35 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® SA8255P	1.562 ms	0 - 35 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® QCS8450	1.756 ms	0 - 60 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	1.432 ms	0 - 16 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	0.459 ms	0 - 33 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® SA7255P	5.061 ms	0 - 33 MB	NPU
GoogLeNet	TFLITE	float	Qualcomm® SA8295P	1.821 ms	0 - 31 MB	NPU
GoogLeNet	TFLITE	float	Snapdragon® 8 Elite Mobile	0.459 ms	0 - 33 MB	NPU
GoogLeNet	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	0.382 ms	0 - 34 MB	NPU
GoogLeNet	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.192 ms	0 - 43 MB	NPU
GoogLeNet	TFLITE	w8a8	Snapdragon® 8 Gen 1 Mobile	0.424 ms	0 - 49 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS6490	1.199 ms	0 - 8 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-8275	0.363 ms	0 - 9 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.241 ms	0 - 2 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® SA8775P	0.424 ms	0 - 30 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® SA8650P	0.424 ms	0 - 30 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® SA8255P	0.424 ms	0 - 30 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® QCS8450	0.424 ms	0 - 49 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-9075	0.327 ms	0 - 9 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-6690	2.248 ms	0 - 29 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-7790	0.329 ms	0 - 29 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-8750	0.161 ms	0 - 28 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® SA7255P	0.835 ms	0 - 28 MB	NPU
GoogLeNet	TFLITE	w8a8	Qualcomm® SA8295P	0.641 ms	0 - 26 MB	NPU
GoogLeNet	TFLITE	w8a8	Snapdragon® 8 Elite Mobile	0.161 ms	0 - 28 MB	NPU
GoogLeNet	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.145 ms	0 - 30 MB	NPU
GoogLeNet	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	0.329 ms	0 - 29 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.