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 1.25.0	Download
ONNX	w8a16	Universal	QAIRT 2.45, ONNX Runtime 1.25.0	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	34.276 ms	212 - 212 MB	NPU
ESRGAN	ONNX	float	Snapdragon® X Elite	65.484 ms	149 - 149 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Gen 3 Mobile	50.833 ms	8 - 744 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Gen 1 Mobile	121.055 ms	3 - 729 MB	NPU
ESRGAN	ONNX	float	Qualcomm® QCS8550 (Proxy)	64.863 ms	7 - 10 MB	NPU
ESRGAN	ONNX	float	Qualcomm® QCS8450	121.055 ms	3 - 729 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Elite Mobile	38.156 ms	6 - 353 MB	NPU
ESRGAN	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	27.615 ms	6 - 358 MB	NPU
ESRGAN	ONNX	float	Qualcomm® QCS9075	107.039 ms	7 - 52 MB	NPU
ESRGAN	ONNX	float	Qualcomm® QCS8750	38.156 ms	6 - 353 MB	NPU
ESRGAN	ONNX	float	Qualcomm® QCS7181	65.484 ms	149 - 149 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® X2 Elite	22.011 ms	212 - 212 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® X Elite	43.602 ms	149 - 149 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	32.036 ms	3 - 1120 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	84.74 ms	0 - 1215 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® QCS6490	244.057 ms	3 - 48 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	42.22 ms	0 - 32 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® QCS8450	84.74 ms	0 - 1215 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® QCS9075	45.393 ms	3 - 48 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	17.336 ms	3 - 1087 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® 8 Elite Mobile	26.611 ms	3 - 923 MB	NPU
ESRGAN	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	90.31 ms	5 - 782 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® QCM6690	1098.627 ms	4 - 723 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® QCS7790	90.31 ms	5 - 782 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® QCS8750	26.611 ms	3 - 923 MB	NPU
ESRGAN	ONNX	w8a16	Qualcomm® QCS7181	43.602 ms	149 - 149 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® X2 Elite	34.377 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® X Elite	64.985 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	49.108 ms	0 - 704 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	122.704 ms	0 - 710 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® QCS8275	451.991 ms	1 - 352 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	65.679 ms	0 - 636 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® QCS8450	122.704 ms	0 - 710 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Elite Mobile	37.943 ms	0 - 334 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA8295P	111.388 ms	1 - 359 MB	NPU
ESRGAN	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	25.371 ms	0 - 334 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® SA7255P	451.991 ms	1 - 352 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® QCS9075	106.612 ms	0 - 5 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® QCS8750	37.943 ms	0 - 334 MB	NPU
ESRGAN	QNN_DLC	float	Qualcomm® QCS7181	64.985 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® X2 Elite	21.919 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® X Elite	43.142 ms	0 - 0 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	31.332 ms	0 - 1038 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® 8 Gen 1 Mobile	77.666 ms	1 - 1114 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® QCS6490	241.622 ms	0 - 3 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® QCS8275	132.941 ms	1 - 672 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	41.621 ms	0 - 22 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® QCS8450	77.666 ms	1 - 1114 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® QCS9075	43.985 ms	0 - 3 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	16.041 ms	0 - 1004 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	27.95 ms	0 - 843 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® SA8295P	65.098 ms	0 - 717 MB	NPU
ESRGAN	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	89.28 ms	0 - 695 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® QCM6690	1115.136 ms	0 - 642 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® SA7255P	132.941 ms	1 - 672 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® QCS7790	89.28 ms	0 - 695 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® QCS8750	27.95 ms	0 - 843 MB	NPU
ESRGAN	QNN_DLC	w8a16	Qualcomm® QCS7181	43.142 ms	0 - 0 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	48.867 ms	3 - 744 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	118.585 ms	0 - 740 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® QCS8275	452.109 ms	4 - 396 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® QCS8550 (Proxy)	62.248 ms	3 - 6 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA8775P	14382.23 ms	3 - 43 MB	GPU
ESRGAN	TFLITE	float	Qualcomm® SA8650P	14382.23 ms	3 - 43 MB	GPU
ESRGAN	TFLITE	float	Qualcomm® SA8255P	14382.23 ms	3 - 43 MB	GPU
ESRGAN	TFLITE	float	Qualcomm® QCS8450	118.585 ms	0 - 740 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Elite Mobile	37.779 ms	3 - 375 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA8295P	111.392 ms	3 - 392 MB	NPU
ESRGAN	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	25.239 ms	3 - 365 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® SA7255P	452.109 ms	4 - 396 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® QCS9075	108.259 ms	3 - 46 MB	NPU
ESRGAN	TFLITE	float	Qualcomm® QCS8750	37.779 ms	3 - 375 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.