Hugging Face's logo

qualcomm
/
NAFNet-DeNoise

Image-to-Image
PyTorch
android
Model card Files
xet
 Community
NAFNet-DeNoise
File size: 11,162 Bytes
491645a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
13f226e
491645a
 
 
 
 
 
 
 
 
 
 
 
13f226e
 
 
 
 
491645a
 
 
 
 
 
13f226e
491645a
 
 
 
 
 
13f226e
491645a
 
 
 
 
 
 
 
 
 
 
 
 
 
13f226e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
491645a
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
---
library_name: pytorch
license: other
tags:
- android
pipeline_tag: image-to-image

---

![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_denoise/web-assets/model_demo.png)

# 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](https://github.com/megvii-research/NAFNet.git).
This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.59.0/src/qai_hub_models/models/nafnet_denoise) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).

Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) 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](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_denoise/releases/v0.59.0/nafnet_denoise-onnx-float.zip)
| ONNX | w8a16 | Universal | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_denoise/releases/v0.59.0/nafnet_denoise-onnx-w8a16.zip)
| QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_denoise/releases/v0.59.0/nafnet_denoise-qnn_dlc-float.zip)
| QNN_DLC | w8a16 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_denoise/releases/v0.59.0/nafnet_denoise-qnn_dlc-w8a16.zip)
| TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_denoise/releases/v0.59.0/nafnet_denoise-tflite-float.zip)

For more device-specific assets and performance metrics, visit **[NAFNet-DeNoise on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/nafnet_denoise)**.


### Option 2: Export with Custom Configurations

Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.59.0/src/qai_hub_models/models/nafnet_denoise) 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](https://github.com/qualcomm/ai-hub-models/blob/v0.59.0/src/qai_hub_models/models/nafnet_denoise) 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](https://github.com/megvii-research/NAFNet/blob/main/LICENSE).

## References
* [Simple Baselines for Image Restoration](https://arxiv.org/abs/2204.04676)
* [Source Model Implementation](https://github.com/megvii-research/NAFNet.git)

## Community
* Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
* For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).