EfficientNet-Lite4: Optimized for Qualcomm Devices

EfficientNet-Lite4 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-Lite4 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 EfficientNet-Lite4 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-Lite4 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

Input resolution: 300x300
Model checkpoint: Imagenet
Model size (float): 51.9 MB
Number of parameters: 13.01M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
EfficientNet-Lite4	ONNX	float	Snapdragon® X2 Elite	1.448 ms	1 - 1 MB	NPU
EfficientNet-Lite4	ONNX	float	Snapdragon® X Elite	2.736 ms	29 - 29 MB	NPU
EfficientNet-Lite4	ONNX	float	Snapdragon® 8 Gen 3 Mobile	1.913 ms	0 - 102 MB	NPU
EfficientNet-Lite4	ONNX	float	Snapdragon® 8 Gen 1 Mobile	5.536 ms	0 - 100 MB	NPU
EfficientNet-Lite4	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	4.631 ms	1 - 5 MB	NPU
EfficientNet-Lite4	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	2.623 ms	0 - 33 MB	NPU
EfficientNet-Lite4	ONNX	float	Qualcomm® QCS8450	5.536 ms	0 - 100 MB	NPU
EfficientNet-Lite4	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	4.145 ms	1 - 4 MB	NPU
EfficientNet-Lite4	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	2.736 ms	29 - 29 MB	NPU
EfficientNet-Lite4	ONNX	float	Qualcomm® Dragonwing™ Q-8750	1.499 ms	0 - 55 MB	NPU
EfficientNet-Lite4	ONNX	float	Snapdragon® 8 Elite Mobile	1.499 ms	0 - 55 MB	NPU
EfficientNet-Lite4	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	1.145 ms	0 - 58 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Snapdragon® X2 Elite	0.646 ms	1 - 1 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Snapdragon® X Elite	1.283 ms	15 - 15 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	0.874 ms	0 - 110 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Snapdragon® 8 Gen 1 Mobile	2.077 ms	0 - 117 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Qualcomm® Dragonwing™ QCS6490	5.978 ms	0 - 3 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-8275	1.335 ms	0 - 4 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.24 ms	0 - 18 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Qualcomm® QCS8450	2.077 ms	0 - 117 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-9075	1.422 ms	0 - 3 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Qualcomm® Dragonwing™ IQ-X7181	1.283 ms	15 - 15 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Qualcomm® Dragonwing™ Q-6690	10.939 ms	0 - 194 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Qualcomm® Dragonwing™ Q-7790	1.592 ms	0 - 72 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Qualcomm® Dragonwing™ Q-8750	0.715 ms	0 - 73 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Snapdragon® 8 Elite Mobile	0.715 ms	0 - 73 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.601 ms	0 - 73 MB	NPU
EfficientNet-Lite4	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	1.592 ms	0 - 72 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Snapdragon® X2 Elite	1.878 ms	1 - 1 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Snapdragon® X Elite	3.486 ms	1 - 1 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	2.209 ms	0 - 98 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	6.89 ms	0 - 99 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	4.841 ms	1 - 4 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	3.16 ms	1 - 3 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® SA8775P	4.785 ms	1 - 55 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® SA8650P	4.785 ms	1 - 55 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® SA8255P	4.785 ms	1 - 55 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® QCS8450	6.89 ms	0 - 99 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	4.512 ms	1 - 4 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	3.486 ms	1 - 1 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	1.689 ms	1 - 57 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® SA7255P	15.278 ms	1 - 53 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Qualcomm® SA8295P	6.462 ms	1 - 50 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Snapdragon® 8 Elite Mobile	1.689 ms	1 - 57 MB	NPU
EfficientNet-Lite4	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	1.267 ms	1 - 58 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Snapdragon® X2 Elite	0.734 ms	0 - 0 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Snapdragon® X Elite	1.444 ms	0 - 0 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	0.907 ms	0 - 98 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Snapdragon® 8 Gen 1 Mobile	2.352 ms	0 - 107 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS6490	7.183 ms	0 - 2 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-8275	1.329 ms	0 - 3 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	1.263 ms	0 - 95 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® SA8775P	1.611 ms	0 - 66 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® SA8650P	1.611 ms	0 - 66 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® SA8255P	1.611 ms	0 - 66 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® QCS8450	2.352 ms	0 - 107 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-9075	1.419 ms	0 - 2 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® Dragonwing™ IQ-X7181	1.444 ms	0 - 0 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-6690	11.29 ms	0 - 183 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-7790	1.614 ms	0 - 65 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® Dragonwing™ Q-8750	0.662 ms	0 - 63 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® SA7255P	3.364 ms	0 - 63 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Qualcomm® SA8295P	2.532 ms	0 - 61 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Snapdragon® 8 Elite Mobile	0.662 ms	0 - 63 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.526 ms	0 - 67 MB	NPU
EfficientNet-Lite4	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	1.614 ms	0 - 65 MB	NPU
EfficientNet-Lite4	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	2.198 ms	0 - 112 MB	NPU
EfficientNet-Lite4	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	6.673 ms	0 - 112 MB	NPU
EfficientNet-Lite4	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	4.852 ms	0 - 33 MB	NPU
EfficientNet-Lite4	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	3.128 ms	0 - 3 MB	NPU
EfficientNet-Lite4	TFLITE	float	Qualcomm® SA8775P	4.792 ms	0 - 66 MB	NPU
EfficientNet-Lite4	TFLITE	float	Qualcomm® SA8650P	4.792 ms	0 - 66 MB	NPU
EfficientNet-Lite4	TFLITE	float	Qualcomm® SA8255P	4.792 ms	0 - 66 MB	NPU
EfficientNet-Lite4	TFLITE	float	Qualcomm® QCS8450	6.673 ms	0 - 112 MB	NPU
EfficientNet-Lite4	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	4.539 ms	0 - 32 MB	NPU
EfficientNet-Lite4	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	1.689 ms	0 - 67 MB	NPU
EfficientNet-Lite4	TFLITE	float	Qualcomm® SA7255P	15.322 ms	0 - 65 MB	NPU
EfficientNet-Lite4	TFLITE	float	Qualcomm® SA8295P	6.457 ms	0 - 55 MB	NPU
EfficientNet-Lite4	TFLITE	float	Snapdragon® 8 Elite Mobile	1.689 ms	0 - 67 MB	NPU
EfficientNet-Lite4	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	1.26 ms	0 - 68 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	0.657 ms	0 - 98 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Snapdragon® 8 Gen 1 Mobile	1.755 ms	0 - 106 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS6490	5.752 ms	0 - 31 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-8275	1.046 ms	0 - 18 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® Dragonwing™ QCS8550 (Proxy)	0.917 ms	0 - 5 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® SA8775P	1.278 ms	0 - 63 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® SA8650P	1.278 ms	0 - 63 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® SA8255P	1.278 ms	0 - 63 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® QCS8450	1.755 ms	0 - 106 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® Dragonwing™ IQ-9075	1.069 ms	0 - 17 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-6690	10.033 ms	0 - 178 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-7790	1.246 ms	0 - 175 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® Dragonwing™ Q-8750	0.52 ms	0 - 57 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® SA7255P	2.762 ms	0 - 59 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Qualcomm® SA8295P	2.14 ms	0 - 56 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Snapdragon® 8 Elite Mobile	0.52 ms	0 - 57 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	0.415 ms	0 - 62 MB	NPU
EfficientNet-Lite4	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	1.246 ms	0 - 175 MB	NPU

License

The license for the original implementation of EfficientNet-Lite4 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-Lite4

EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks

Paper • 1905.11946 • Published May 28, 2019 • 4