NAFNet-DeNoise: Optimized for Qualcomm Devices

NAFNET is designed for lightweight real-time denoising of images.

This is based on the implementation of NAFNet-DeNoise 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 NAFNet-DeNoise 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 NAFNet-DeNoise on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_editing

Model Stats:

• Model checkpoint: NAFNet-SIDD-width64
• Input resolution: 256x256
• Number of parameters: 115.98M
• Model size (float): 463.93 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
NAFNet-DeNoise	ONNX	float	Snapdragon® X2 Elite	14.988 ms	4 - 4 MB	NPU
NAFNet-DeNoise	ONNX	float	Snapdragon® X Elite	33.919 ms	228 - 228 MB	NPU
NAFNet-DeNoise	ONNX	float	Snapdragon® 8 Gen 3 Mobile	27.636 ms	2 - 851 MB	NPU
NAFNet-DeNoise	ONNX	float	Snapdragon® 8 Gen 1 Mobile	61.4 ms	3 - 722 MB	NPU
NAFNet-DeNoise	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	34.841 ms	0 - 258 MB	NPU
NAFNet-DeNoise	ONNX	float	Qualcomm® QCS8450	61.4 ms	3 - 722 MB	NPU
NAFNet-DeNoise	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	43.039 ms	2 - 5 MB	NPU
NAFNet-DeNoise	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	33.919 ms	228 - 228 MB	NPU
NAFNet-DeNoise	ONNX	float	Qualcomm® Dragonwing™ Q-8750	19.714 ms	2 - 498 MB	NPU
NAFNet-DeNoise	ONNX	float	Snapdragon® 8 Elite Mobile	19.714 ms	2 - 498 MB	NPU
NAFNet-DeNoise	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	14.475 ms	0 - 535 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Snapdragon® X2 Elite	14.577 ms	1 - 1 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Snapdragon® X Elite	33.188 ms	120 - 120 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	24.378 ms	1 - 697 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Snapdragon® 8 Gen 1 Mobile	53.664 ms	2 - 710 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Qualcomm® Dragonwing™ QCS6490	144.383 ms	1 - 4 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	32.729 ms	0 - 130 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Qualcomm® QCS8450	53.664 ms	2 - 710 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-9075	32.725 ms	1 - 4 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Qualcomm® Dragonwing™ IQ-X7181	33.188 ms	120 - 120 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Qualcomm® Dragonwing™ Q-8750	19.222 ms	1 - 566 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Snapdragon® 8 Elite Mobile	19.222 ms	1 - 566 MB	NPU
NAFNet-DeNoise	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	13.027 ms	0 - 684 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Snapdragon® X2 Elite	15.695 ms	1 - 1 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Snapdragon® X Elite	35.856 ms	1 - 1 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	28.778 ms	1 - 892 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	60.876 ms	0 - 728 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8275	136.431 ms	1 - 546 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	38.207 ms	1 - 19 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® SA8775P	43.544 ms	1 - 559 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® SA8650P	43.544 ms	1 - 559 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® SA8255P	43.544 ms	1 - 559 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® QCS8450	60.876 ms	0 - 728 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	48.478 ms	1 - 4 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	35.856 ms	1 - 1 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	20.539 ms	1 - 542 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® SA7255P	136.431 ms	1 - 546 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Qualcomm® SA8295P	48.018 ms	1 - 399 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Snapdragon® 8 Elite Mobile	20.539 ms	1 - 542 MB	NPU
NAFNet-DeNoise	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	14.726 ms	0 - 621 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Snapdragon® X2 Elite	15.063 ms	0 - 0 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Snapdragon® X Elite	33.124 ms	0 - 0 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	22.473 ms	0 - 687 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Snapdragon® 8 Gen 1 Mobile	57.732 ms	0 - 719 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS6490	141.888 ms	0 - 3 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8275	65.077 ms	1 - 576 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® Dragonwing™ QCS8550 (Proxy)	31.599 ms	0 - 3 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® SA8775P	31.638 ms	1 - 576 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® SA8650P	31.638 ms	1 - 576 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® SA8255P	31.638 ms	1 - 576 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® QCS8450	57.732 ms	0 - 719 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-9075	32.191 ms	2 - 5 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® Dragonwing™ IQ-X7181	33.124 ms	0 - 0 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-6690	354.702 ms	0 - 954 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-7790	39.185 ms	0 - 813 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® Dragonwing™ Q-8750	21.24 ms	0 - 584 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® SA7255P	65.077 ms	1 - 576 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Qualcomm® SA8295P	40.513 ms	0 - 574 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Snapdragon® 8 Elite Mobile	21.24 ms	0 - 584 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	13.267 ms	0 - 706 MB	NPU
NAFNet-DeNoise	QNN_DLC	w8a16	Snapdragon® 7 Gen 4 Mobile	39.185 ms	0 - 813 MB	NPU
NAFNet-DeNoise	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	28.866 ms	1 - 1125 MB	NPU
NAFNet-DeNoise	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	64.27 ms	1 - 962 MB	NPU
NAFNet-DeNoise	TFLITE	float	Qualcomm® Dragonwing™ QCS8275	136.679 ms	1 - 765 MB	NPU
NAFNet-DeNoise	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	38.107 ms	1 - 4 MB	NPU
NAFNet-DeNoise	TFLITE	float	Qualcomm® SA8775P	43.637 ms	1 - 777 MB	NPU
NAFNet-DeNoise	TFLITE	float	Qualcomm® SA8650P	43.637 ms	1 - 777 MB	NPU
NAFNet-DeNoise	TFLITE	float	Qualcomm® SA8255P	43.637 ms	1 - 777 MB	NPU
NAFNet-DeNoise	TFLITE	float	Qualcomm® QCS8450	64.27 ms	1 - 962 MB	NPU
NAFNet-DeNoise	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	44.226 ms	1 - 233 MB	NPU
NAFNet-DeNoise	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	20.402 ms	1 - 764 MB	NPU
NAFNet-DeNoise	TFLITE	float	Qualcomm® SA7255P	136.679 ms	1 - 765 MB	NPU
NAFNet-DeNoise	TFLITE	float	Qualcomm® SA8295P	50.795 ms	1 - 629 MB	NPU
NAFNet-DeNoise	TFLITE	float	Snapdragon® 8 Elite Mobile	20.402 ms	1 - 764 MB	NPU
NAFNet-DeNoise	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	14.787 ms	1 - 849 MB	NPU

License

• The license for the original implementation of NAFNet-DeNoise can be found here.

References

• Simple Baselines for Image Restoration
• 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.