EfficientNet-B4: Optimized for Qualcomm Devices

EfficientNetB4 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 EfficientNet-B4 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 2.50, ONNX Runtime 1.27.1	Download
ONNX	w8a16	Universal	QAIRT 2.50, ONNX Runtime 1.27.1	Download
QNN_DLC	float	Universal	QAIRT 2.50	Download
QNN_DLC	w8a16	Universal	QAIRT 2.50	Download
TFLITE	float	Universal	QAIRT 2.50	Download

For more device-specific assets and performance metrics, visit EfficientNet-B4 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 EfficientNet-B4 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

Input resolution: 380x380
Model checkpoint: Imagenet
Model size (float): 73.6 MB
Model size (w8a16): 24.0 MB
Number of parameters: 19.3M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
EfficientNet-B4	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	3.233 ms	0 - 200 MB	NPU
EfficientNet-B4	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	4.085 ms	0 - 90 MB	NPU
EfficientNet-B4	ONNX	float	Snapdragon® X2 Elite	4.061 ms	2 - 2 MB	NPU
EfficientNet-B4	ONNX	float	Snapdragon® X Elite	7.624 ms	45 - 45 MB	NPU
EfficientNet-B4	ONNX	float	Snapdragon® 8 Gen 3 Mobile	5.291 ms	0 - 147 MB	NPU
EfficientNet-B4	ONNX	float	Snapdragon® 8 Gen 1 Mobile	17.255 ms	2 - 190 MB	NPU
EfficientNet-B4	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	9.563 ms	2 - 7 MB	NPU
EfficientNet-B4	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	7.241 ms	0 - 48 MB	NPU
EfficientNet-B4	ONNX	float	Qualcomm® QCS8450	17.255 ms	2 - 190 MB	NPU
EfficientNet-B4	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	9.31 ms	1 - 6 MB	NPU
EfficientNet-B4	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	7.624 ms	45 - 45 MB	NPU
EfficientNet-B4	ONNX	float	Qualcomm® Dragonwing™ Q-8750	4.085 ms	0 - 90 MB	NPU
EfficientNet-B4	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	2.773 ms	0 - 176 MB	NPU
EfficientNet-B4	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	3.444 ms	0 - 168 MB	NPU
EfficientNet-B4	ONNX	w8a16	Snapdragon® X2 Elite	2.91 ms	2 - 2 MB	NPU
EfficientNet-B4	ONNX	w8a16	Snapdragon® X Elite	8.115 ms	24 - 24 MB	NPU
EfficientNet-B4	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	5.158 ms	1 - 227 MB	NPU
EfficientNet-B4	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	8.668 ms	1 - 227 MB	NPU
EfficientNet-B4	ONNX	w8a16	Qualcomm® Dragonwing™ QCS6490	25.189 ms	0 - 4 MB	NPU
EfficientNet-B4	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-8275	6.972 ms	1 - 5 MB	NPU
EfficientNet-B4	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	7.726 ms	0 - 184 MB	NPU
EfficientNet-B4	ONNX	w8a16	Qualcomm® QCS8450	8.668 ms	1 - 227 MB	NPU
EfficientNet-B4	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	8.071 ms	1 - 4 MB	NPU
EfficientNet-B4	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	8.115 ms	24 - 24 MB	NPU
EfficientNet-B4	ONNX	w8a16	Qualcomm® Dragonwing™ Q-6690	43.05 ms	0 - 288 MB	NPU
EfficientNet-B4	ONNX	w8a16	Qualcomm® Dragonwing™ Q-7790	8.991 ms	1 - 281 MB	NPU
EfficientNet-B4	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	3.444 ms	0 - 168 MB	NPU
EfficientNet-B4	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	8.991 ms	1 - 281 MB	NPU
EfficientNet-B4	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	3.238 ms	2 - 197 MB	NPU
EfficientNet-B4	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	4.311 ms	0 - 84 MB	NPU
EfficientNet-B4	QNN_DLC	float	Snapdragon® X2 Elite	4.404 ms	2 - 2 MB	NPU
EfficientNet-B4	QNN_DLC	float	Snapdragon® X Elite	8.712 ms	2 - 2 MB	NPU
EfficientNet-B4	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	5.826 ms	0 - 143 MB	NPU
EfficientNet-B4	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	19.907 ms	0 - 182 MB	NPU
EfficientNet-B4	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	10.02 ms	2 - 6 MB	NPU
EfficientNet-B4	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	8.065 ms	2 - 3 MB	NPU
EfficientNet-B4	QNN_DLC	float	Qualcomm® SA8775P	10.305 ms	2 - 84 MB	NPU
EfficientNet-B4	QNN_DLC	float	Qualcomm® SA8650P	10.305 ms	2 - 84 MB	NPU
EfficientNet-B4	QNN_DLC	float	Qualcomm® SA8255P	10.305 ms	2 - 84 MB	NPU
EfficientNet-B4	QNN_DLC	float	Qualcomm® QCS8450	19.907 ms	0 - 182 MB	NPU
EfficientNet-B4	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	10.045 ms	2 - 5 MB	NPU
EfficientNet-B4	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	8.712 ms	2 - 2 MB	NPU
EfficientNet-B4	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	4.311 ms	0 - 84 MB	NPU
EfficientNet-B4	QNN_DLC	float	Qualcomm® SA8295P	18.599 ms	2 - 123 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	3.039 ms	1 - 158 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	3.763 ms	0 - 151 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Snapdragon® X2 Elite	3.569 ms	1 - 1 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Snapdragon® X Elite	9.146 ms	1 - 1 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	5.63 ms	1 - 201 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Snapdragon® 8 Gen 1 Mobile	10.419 ms	1 - 205 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS6490	27.248 ms	1 - 4 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-8275	7.622 ms	1 - 5 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	8.466 ms	1 - 2 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® SA8775P	8.943 ms	1 - 147 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® SA8650P	8.943 ms	1 - 147 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® SA8255P	8.943 ms	1 - 147 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® QCS8450	10.419 ms	1 - 205 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	8.74 ms	1 - 4 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	9.146 ms	1 - 1 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-6690	44.793 ms	1 - 271 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-7790	9.7 ms	1 - 265 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	3.763 ms	0 - 151 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® SA7255P	15.449 ms	1 - 144 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Qualcomm® SA8295P	10.899 ms	1 - 147 MB	NPU
EfficientNet-B4	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	9.7 ms	1 - 265 MB	NPU
EfficientNet-B4	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	3.23 ms	0 - 99 MB	NPU
EfficientNet-B4	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	4.308 ms	0 - 100 MB	NPU
EfficientNet-B4	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	5.773 ms	0 - 164 MB	NPU
EfficientNet-B4	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	20.039 ms	0 - 203 MB	NPU
EfficientNet-B4	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	10.013 ms	0 - 50 MB	NPU
EfficientNet-B4	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	8.007 ms	0 - 3 MB	NPU
EfficientNet-B4	TFLITE	float	Qualcomm® QCS8450	20.039 ms	0 - 203 MB	NPU
EfficientNet-B4	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	9.932 ms	0 - 50 MB	NPU
EfficientNet-B4	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	4.308 ms	0 - 100 MB	NPU
EfficientNet-B4	TFLITE	float	Qualcomm® SA8295P	18.598 ms	0 - 139 MB	NPU

License

The license for the original implementation of EfficientNet-B4 can be found here.

References

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.

Downloads last month: -; Downloads are not tracked for this model. How to track

Paper for qualcomm/EfficientNet-B4

EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks

Paper • 1905.11946 • Published May 28, 2019 • 4