ESRGAN: Optimized for Qualcomm Devices

ESRGAN is a machine learning model that upscales an image with minimal loss in quality.

This is based on the implementation of ESRGAN 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 1.30.0	Download
QNN_DLC	float	Universal	QAIRT 2.50	Download
TFLITE	float	Universal	QAIRT 2.50	Download

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

Model Details

Model Type: Model_use_case.super_resolution

Model Stats:

• Input resolution: 128x128
• Model checkpoint: ESRGAN_x4
• Model size (float): 63.9 MB
• Number of parameters: 16.7M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
ESRGAN	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	26.852 ms	5 - 337 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	38.67 ms	6 - 330 MB	NPU
ESRGAN	ONNX	float	Snapdragon® X2 Elite	34.433 ms	8 - 8 MB	NPU
ESRGAN	ONNX	float	Snapdragon® X Elite	65.297 ms	38 - 38 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Gen 3 Mobile	49.302 ms	0 - 719 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Gen 1 Mobile	113.948 ms	3 - 715 MB	NPU
ESRGAN	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	127.242 ms	6 - 11 MB	NPU
ESRGAN	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	67.123 ms	0 - 43 MB	NPU
ESRGAN	ONNX	float	Qualcomm® QCS8450	113.948 ms	3 - 715 MB	NPU
ESRGAN	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	105.762 ms	8 - 11 MB	NPU
ESRGAN	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	65.297 ms	38 - 38 MB	NPU
ESRGAN	ONNX	float	Qualcomm® Dragonwing™ Q-8750	38.67 ms	6 - 330 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	24.723 ms	0 - 313 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	37.922 ms	0 - 307 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® X2 Elite	34.399 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® X Elite	64.815 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	49.803 ms	0 - 674 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	121.567 ms	0 - 671 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	126.676 ms	0 - 6 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	62.782 ms	0 - 3 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA8775P	105.12 ms	1 - 323 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA8650P	105.12 ms	1 - 323 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA8255P	105.12 ms	1 - 323 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® QCS8450	121.567 ms	0 - 671 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	105.409 ms	0 - 6 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	64.815 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	37.922 ms	0 - 307 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA8295P	110.072 ms	0 - 333 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	25.75 ms	3 - 354 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	37.897 ms	3 - 347 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	48.453 ms	39 - 756 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	137.693 ms	3 - 709 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	126.779 ms	3 - 47 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	66.487 ms	3 - 6 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA8775P	105.211 ms	3 - 367 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA8650P	105.211 ms	3 - 367 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA8255P	105.211 ms	3 - 367 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® QCS8450	137.693 ms	3 - 709 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	105.822 ms	3 - 46 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	37.897 ms	3 - 347 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA8295P	110.112 ms	3 - 370 MB	NPU

License

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

References

• ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks
• 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.