Hugging Face's logo

qualcomm
/
NAFNet-DeBlur

Image-to-Image
PyTorch
android
Model card Files
xet
 Community
NAFNet-DeBlur
File size: 15,418 Bytes
83cdeb0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d507ebb
83cdeb0
 
 
 
 
 
 
 
 
 
 
 
d507ebb
 
 
 
 
83cdeb0
 
 
 
 
 
d507ebb
83cdeb0
 
 
 
 
 
d507ebb
83cdeb0
 
 
 
 
 
 
 
 
 
 
 
 
 
d507ebb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
83cdeb0
 
 
 
 
 
 
 
 
 
 
 
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
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
---
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_deblur/web-assets/model_demo.png)

# NAFNet-DeBlur: Optimized for Qualcomm Devices

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

This is based on the implementation of NAFNet-DeBlur 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/v0.58.0/src/qai_hub_models/models/nafnet_deblur) 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 1.25.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_deblur/releases/v0.58.0/nafnet_deblur-onnx-float.zip)
| ONNX | w8a16 | Universal | QAIRT 2.45, ONNX Runtime 1.25.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_deblur/releases/v0.58.0/nafnet_deblur-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_deblur/releases/v0.58.0/nafnet_deblur-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_deblur/releases/v0.58.0/nafnet_deblur-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_deblur/releases/v0.58.0/nafnet_deblur-tflite-float.zip)

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


### Option 2: Export with Custom Configurations

Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.58.0/src/qai_hub_models/models/nafnet_deblur) 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-DeBlur on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.58.0/src/qai_hub_models/models/nafnet_deblur) for usage instructions.

## Model Details

**Model Type:** Model_use_case.image_editing

**Model Stats:**
- Model checkpoint: NAFNet-REDS-width64
- Input resolution: 360x640
- Number of parameters: 67.89M
- Model size (float): 271.55 MB

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| NAFNet-DeBlur | ONNX | float | Snapdragon® X2 Elite | 47.739 ms | 11 - 11 MB | NPU
| NAFNet-DeBlur | ONNX | float | Snapdragon® X Elite | 113.074 ms | 142 - 142 MB | NPU
| NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 81.438 ms | 0 - 1195 MB | NPU
| NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 183.315 ms | 10 - 862 MB | NPU
| NAFNet-DeBlur | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 112.679 ms | 0 - 161 MB | NPU
| NAFNet-DeBlur | ONNX | float | Qualcomm® QCS8450 | 183.315 ms | 10 - 862 MB | NPU
| NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 45.045 ms | 7 - 995 MB | NPU
| NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Elite Mobile | 63.561 ms | 8 - 941 MB | NPU
| NAFNet-DeBlur | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 63.561 ms | 8 - 941 MB | NPU
| NAFNet-DeBlur | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 113.074 ms | 142 - 142 MB | NPU
| NAFNet-DeBlur | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 152.496 ms | 9 - 54 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® X2 Elite | 43.843 ms | 5 - 5 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® X Elite | 92.128 ms | 77 - 77 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 64.767 ms | 4 - 1084 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 89.707 ms | 0 - 84 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 54.509 ms | 4 - 967 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 105.292 ms | 3 - 7 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 133.706 ms | 6 - 1140 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 39.539 ms | 2 - 992 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 1326.727 ms | 0 - 1528 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 133.706 ms | 6 - 1140 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 54.509 ms | 4 - 967 MB | NPU
| NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 92.128 ms | 77 - 77 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® X2 Elite | 19.589 ms | 4 - 4 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® X Elite | 46.725 ms | 74 - 74 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 30.415 ms | 0 - 813 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® 8 Gen 1 Mobile | 85.307 ms | 3 - 848 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 223.392 ms | 1 - 5 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 44.802 ms | 0 - 83 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® QCS8450 | 85.307 ms | 3 - 848 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® 8 Elite Mobile | 26.793 ms | 0 - 598 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 59.482 ms | 2 - 835 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 18.968 ms | 0 - 748 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 515.518 ms | 0 - 941 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 59.482 ms | 2 - 835 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 26.793 ms | 0 - 598 MB | NPU
| NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ IQ-X7181 | 46.725 ms | 74 - 74 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Snapdragon® X2 Elite | 47.531 ms | 3 - 3 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Snapdragon® X Elite | 112.583 ms | 3 - 3 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 81.126 ms | 3 - 1202 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 185.051 ms | 2 - 839 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® QCS8275 | 483.078 ms | 2 - 913 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 109.221 ms | 3 - 6 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA8775P | 139.95 ms | 1 - 917 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA8650P | 139.95 ms | 1 - 917 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA8255P | 139.95 ms | 1 - 917 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® QCS8450 | 185.051 ms | 2 - 839 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 44.648 ms | 0 - 1003 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 63.005 ms | 0 - 948 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA7255P | 483.078 ms | 2 - 913 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA8295P | 162.305 ms | 0 - 632 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 63.005 ms | 0 - 948 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 112.583 ms | 3 - 3 MB | NPU
| NAFNet-DeBlur | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 152.687 ms | 3 - 10 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 43.227 ms | 1 - 1 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® X Elite | 91.189 ms | 1 - 1 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 62.713 ms | 0 - 1032 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® QCS8275 | 192.769 ms | 2 - 780 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 88.626 ms | 2 - 5 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® SA8775P | 86.069 ms | 2 - 780 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® SA8650P | 86.069 ms | 2 - 780 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® SA8255P | 86.069 ms | 2 - 780 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 8 Elite Mobile | 52.175 ms | 1 - 934 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 85.581 ms | 3 - 7 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® SA7255P | 192.769 ms | 2 - 780 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 136.397 ms | 1 - 1190 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 39.251 ms | 1 - 1044 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 1294.492 ms | 1 - 1486 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 136.397 ms | 1 - 1190 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 52.175 ms | 1 - 934 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 91.189 ms | 1 - 1 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 21.311 ms | 1 - 1 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® X Elite | 47.39 ms | 1 - 1 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 30.401 ms | 1 - 786 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® 8 Gen 1 Mobile | 85.808 ms | 1 - 812 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 217.601 ms | 1 - 4 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® QCS8275 | 91.145 ms | 1 - 605 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 45.054 ms | 1 - 68 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® SA8775P | 44.383 ms | 1 - 605 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® SA8650P | 44.383 ms | 1 - 605 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® SA8255P | 44.383 ms | 1 - 605 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® QCS8450 | 85.808 ms | 1 - 812 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® 8 Elite Mobile | 27.487 ms | 1 - 602 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 56.27 ms | 1 - 827 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 44.56 ms | 1 - 4 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® SA7255P | 91.145 ms | 1 - 605 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® SA8295P | 62.501 ms | 1 - 603 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 17.199 ms | 1 - 695 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 493.208 ms | 1 - 963 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 56.27 ms | 1 - 827 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 27.487 ms | 1 - 602 MB | NPU
| NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-X7181 | 47.39 ms | 1 - 1 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 80.988 ms | 3 - 1292 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 210.907 ms | 4 - 1008 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Qualcomm® QCS8275 | 483.041 ms | 3 - 1001 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 110.539 ms | 3 - 7 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8775P | 139.99 ms | 3 - 1005 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8650P | 139.99 ms | 3 - 1005 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8255P | 139.99 ms | 3 - 1005 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Qualcomm® QCS8450 | 210.907 ms | 4 - 1008 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 44.656 ms | 3 - 1090 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Elite Mobile | 62.301 ms | 2 - 1032 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Qualcomm® SA7255P | 483.041 ms | 3 - 1001 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8295P | 162.41 ms | 3 - 738 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 62.301 ms | 2 - 1032 MB | NPU
| NAFNet-DeBlur | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 152.356 ms | 1 - 150 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 60.273 ms | 1 - 1099 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Snapdragon® 8 Gen 1 Mobile | 123.385 ms | 1 - 1192 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® QCS8275 | 169.472 ms | 1 - 868 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 84.724 ms | 1 - 5 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® SA8775P | 84.58 ms | 1 - 868 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® SA8650P | 84.58 ms | 1 - 868 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® SA8255P | 84.58 ms | 1 - 868 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® QCS8450 | 123.385 ms | 1 - 1192 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Snapdragon® 8 Elite Mobile | 40.239 ms | 1 - 703 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 885.229 ms | 138 - 542 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 109.926 ms | 1 - 81 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® SA7255P | 169.472 ms | 1 - 868 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® SA8295P | 116.821 ms | 1 - 866 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 31.605 ms | 1 - 1209 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 885.229 ms | 138 - 542 MB | NPU
| NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 40.239 ms | 1 - 703 MB | NPU

## License
* The license for the original implementation of NAFNet-DeBlur 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).