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 2.45, ONNX Runtime 1.27.1	Download
ONNX	w8a16	Universal	QAIRT 2.45, ONNX Runtime 1.27.1	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
QNN_DLC	w8a16	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	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:

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

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
ESRGAN	ONNX	float	Snapdragon® X2 Elite	34.263 ms	9 - 9 MB	NPU
ESRGAN	ONNX	float	Snapdragon® X Elite	65.491 ms	38 - 38 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Gen 3 Mobile	50.456 ms	0 - 745 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Gen 1 Mobile	126.988 ms	4 - 739 MB	NPU
ESRGAN	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	69.332 ms	7 - 27 MB	NPU
ESRGAN	ONNX	float	Qualcomm® QCS8450	126.988 ms	4 - 739 MB	NPU
ESRGAN	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	106.674 ms	7 - 10 MB	NPU
ESRGAN	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	65.491 ms	38 - 38 MB	NPU
ESRGAN	ONNX	float	Qualcomm® Dragonwing™ Q-8750	37.998 ms	7 - 356 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Elite Mobile	37.998 ms	7 - 356 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	27.319 ms	5 - 360 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® X2 Elite	21.971 ms	4 - 4 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® X Elite	43.486 ms	26 - 26 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	32.193 ms	3 - 1130 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	86.206 ms	4 - 1215 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® Dragonwing™ QCS6490	323.469 ms	4 - 7 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	42.264 ms	0 - 31 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® QCS8450	86.206 ms	4 - 1215 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	37.762 ms	3 - 6 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	43.486 ms	26 - 26 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	26.581 ms	3 - 924 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® 8 Elite Mobile	26.581 ms	3 - 924 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	16.833 ms	1 - 1088 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® X2 Elite	34.365 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® X Elite	64.984 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	50.384 ms	0 - 705 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	132.45 ms	0 - 711 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8275	452.066 ms	1 - 352 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	68.068 ms	0 - 8 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA8775P	105.473 ms	1 - 352 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA8650P	105.473 ms	1 - 352 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA8255P	105.473 ms	1 - 352 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® QCS8450	132.45 ms	0 - 711 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	106.113 ms	0 - 5 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	64.984 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	37.857 ms	0 - 333 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA7255P	452.066 ms	1 - 352 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA8295P	111.345 ms	0 - 358 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Elite Mobile	37.857 ms	0 - 333 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	25.598 ms	0 - 335 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® X2 Elite	21.894 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® X Elite	43.112 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	31.493 ms	0 - 1036 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® 8 Gen 1 Mobile	80.42 ms	0 - 1115 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS6490	320.28 ms	0 - 3 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8275	132.863 ms	1 - 671 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	42.194 ms	0 - 3 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® SA8775P	37.639 ms	0 - 673 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® SA8650P	37.639 ms	0 - 673 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® SA8255P	37.639 ms	0 - 673 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® QCS8450	80.42 ms	0 - 1115 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	38.052 ms	2 - 5 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	43.112 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-6690	1117.63 ms	0 - 631 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-7790	88.783 ms	2 - 697 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	28.142 ms	0 - 841 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® SA7255P	132.863 ms	1 - 671 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® SA8295P	65.061 ms	0 - 717 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	28.142 ms	0 - 841 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	16.38 ms	0 - 1004 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	88.783 ms	2 - 697 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	49.307 ms	0 - 740 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	133.76 ms	0 - 740 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® Dragonwing™ QCS8275	452.139 ms	4 - 396 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	67.891 ms	3 - 7 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA8775P	105.556 ms	3 - 396 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA8650P	105.556 ms	3 - 396 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA8255P	105.556 ms	3 - 396 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® QCS8450	133.76 ms	0 - 740 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	105.885 ms	1 - 44 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	37.841 ms	1 - 374 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA7255P	452.139 ms	4 - 396 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA8295P	111.402 ms	3 - 392 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Elite Mobile	37.841 ms	1 - 374 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	25.613 ms	3 - 376 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.