Hugging Face's logo

qualcomm
/
NAFNet-DeBlur

Image-to-Image
PyTorch
android
Model card Files
xet
 Community
qaihm-bot commited on
Commit
d507ebb
·
verified ·
1 Parent(s): ac37b5d

v0.58.0

Browse files

See https://github.com/qualcomm/ai-hub-models/releases/v0.58.0 for changelog.

Files changed (2) hide show
  1. README.md +130 -73
  2. release_assets.json +6 -6
README.md CHANGED
@@ -14,7 +14,7 @@ pipeline_tag: image-to-image
14
  NAFNET is designed for lightweight real-time deblurring of images.
15
 
16
  This is based on the implementation of NAFNet-DeBlur found [here](https://github.com/megvii-research/NAFNet.git).
17
- 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/v0.57.3/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).
18
 
19
  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.
20
 
@@ -27,25 +27,25 @@ Below are pre-exported model assets ready for deployment.
27
 
28
  | Runtime | Precision | Chipset | SDK Versions | Download |
29
  |---|---|---|---|---|
30
- | 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/v0.57.3/nafnet_deblur-onnx-float.zip)
31
- | 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/v0.57.3/nafnet_deblur-onnx-w8a16.zip)
32
- | 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/v0.57.3/nafnet_deblur-qnn_dlc-float.zip)
33
- | 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/v0.57.3/nafnet_deblur-qnn_dlc-w8a16.zip)
34
- | TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_deblur/releases/v0.57.3/nafnet_deblur-tflite-float.zip)
35
 
36
  For more device-specific assets and performance metrics, visit **[NAFNet-DeBlur on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/nafnet_deblur)**.
37
 
38
 
39
  ### Option 2: Export with Custom Configurations
40
 
41
- Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.57.3/src/qai_hub_models/models/nafnet_deblur) Python library to compile and export the model with your own:
42
  - Custom weights (e.g., fine-tuned checkpoints)
43
  - Custom input shapes
44
  - Target device and runtime configurations
45
 
46
  This option is ideal if you need to customize the model beyond the default configuration provided here.
47
 
48
- See our repository for [NAFNet-DeBlur on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.57.3/src/qai_hub_models/models/nafnet_deblur) for usage instructions.
49
 
50
  ## Model Details
51
 
@@ -60,71 +60,128 @@ See our repository for [NAFNet-DeBlur on GitHub](https://github.com/qualcomm/ai-
60
  ## Performance Summary
61
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
62
  |---|---|---|---|---|---|---
63
- | NAFNet-DeBlur | ONNX | float | Snapdragon® X2 Elite | 47.541 ms | 210 - 210 MB | NPU
64
- | NAFNet-DeBlur | ONNX | float | Snapdragon® X Elite | 113.086 ms | 180 - 180 MB | NPU
65
- | NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 82.051 ms | 9 - 1195 MB | NPU
66
- | NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 187.013 ms | 9 - 850 MB | NPU
67
- | NAFNet-DeBlur | ONNX | float | Qualcomm® QCS8550 (Proxy) | 111.158 ms | 0 - 161 MB | NPU
68
- | NAFNet-DeBlur | ONNX | float | Qualcomm® QCS8450 | 187.013 ms | 9 - 850 MB | NPU
69
- | NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Elite Mobile | 63.413 ms | 6 - 937 MB | NPU
70
- | NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 44.652 ms | 6 - 987 MB | NPU
71
- | NAFNet-DeBlur | ONNX | float | Qualcomm® QCS9075 | 152.496 ms | 9 - 54 MB | NPU
72
- | NAFNet-DeBlur | ONNX | float | Qualcomm® QCS8750 | 63.413 ms | 6 - 937 MB | NPU
73
- | NAFNet-DeBlur | ONNX | float | Qualcomm® QCS7181 | 113.086 ms | 180 - 180 MB | NPU
74
- | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® X2 Elite | 43.906 ms | 212 - 212 MB | NPU
75
- | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® X Elite | 92.108 ms | 181 - 181 MB | NPU
76
- | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 64.608 ms | 4 - 1073 MB | NPU
77
- | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 89.707 ms | 0 - 84 MB | NPU
78
- | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® QCS9075 | 102.685 ms | 4 - 49 MB | NPU
79
- | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 39.582 ms | 4 - 990 MB | NPU
80
- | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 54.208 ms | 4 - 967 MB | NPU
81
- | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 130.905 ms | 4 - 1136 MB | NPU
82
- | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® QCM6690 | 1401.45 ms | 2 - 1527 MB | NPU
83
- | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® QCS7790 | 130.905 ms | 4 - 1136 MB | NPU
84
- | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® QCS8750 | 54.208 ms | 4 - 967 MB | NPU
85
- | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® QCS7181 | 92.108 ms | 181 - 181 MB | NPU
86
- | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® X2 Elite | 47.477 ms | 3 - 3 MB | NPU
87
- | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® X Elite | 112.14 ms | 3 - 3 MB | NPU
88
- | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 80.672 ms | 3 - 1197 MB | NPU
89
- | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 186.997 ms | 2 - 840 MB | NPU
90
- | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® QCS8275 | 483.081 ms | 0 - 912 MB | NPU
91
- | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 109.113 ms | 3 - 5 MB | NPU
92
- | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® QCS8450 | 186.997 ms | 2 - 840 MB | NPU
93
- | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 62.652 ms | 0 - 946 MB | NPU
94
- | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA8295P | 162.413 ms | 0 - 632 MB | NPU
95
- | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 44.29 ms | 3 - 1006 MB | NPU
96
- | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA7255P | 483.081 ms | 0 - 912 MB | NPU
97
- | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® QCS9075 | 152.687 ms | 3 - 10 MB | NPU
98
- | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® QCS8750 | 62.652 ms | 0 - 946 MB | NPU
99
- | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® QCS7181 | 112.14 ms | 3 - 3 MB | NPU
100
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 43.373 ms | 1 - 1 MB | NPU
101
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® X Elite | 91.414 ms | 1 - 1 MB | NPU
102
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 62.996 ms | 1 - 1031 MB | NPU
103
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® QCS8275 | 192.736 ms | 2 - 780 MB | NPU
104
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 88.235 ms | 2 - 4 MB | NPU
105
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 101.826 ms | 1 - 6 MB | NPU
106
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 38.978 ms | 0 - 1043 MB | NPU
107
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 8 Elite Mobile | 51.955 ms | 1 - 933 MB | NPU
108
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 136.755 ms | 1 - 1184 MB | NPU
109
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® QCM6690 | 1295.314 ms | 1 - 1486 MB | NPU
110
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® SA7255P | 192.736 ms | 2 - 780 MB | NPU
111
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® QCS7790 | 136.755 ms | 1 - 1184 MB | NPU
112
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® QCS8750 | 51.955 ms | 1 - 933 MB | NPU
113
- | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® QCS7181 | 91.414 ms | 1 - 1 MB | NPU
114
- | NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 80.884 ms | 3 - 1290 MB | NPU
115
- | NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 210.333 ms | 4 - 1008 MB | NPU
116
- | NAFNet-DeBlur | TFLITE | float | Qualcomm® QCS8275 | 482.969 ms | 3 - 1001 MB | NPU
117
- | NAFNet-DeBlur | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 108.194 ms | 0 - 3 MB | NPU
118
- | NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8775P | 998.477 ms | 0 - 44 MB | GPU
119
- | NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8650P | 998.477 ms | 0 - 44 MB | GPU
120
- | NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8255P | 998.477 ms | 0 - 44 MB | GPU
121
- | NAFNet-DeBlur | TFLITE | float | Qualcomm® QCS8450 | 210.333 ms | 4 - 1008 MB | NPU
122
- | NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Elite Mobile | 63.163 ms | 2 - 1035 MB | NPU
123
- | NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8295P | 162.459 ms | 3 - 738 MB | NPU
124
- | NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 45.053 ms | 2 - 1089 MB | NPU
125
- | NAFNet-DeBlur | TFLITE | float | Qualcomm® SA7255P | 482.969 ms | 3 - 1001 MB | NPU
126
- | NAFNet-DeBlur | TFLITE | float | Qualcomm® QCS9075 | 152.356 ms | 1 - 150 MB | NPU
127
- | NAFNet-DeBlur | TFLITE | float | Qualcomm® QCS8750 | 63.163 ms | 2 - 1035 MB | NPU
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
128
 
129
  ## License
130
  * The license for the original implementation of NAFNet-DeBlur can be found
 
14
  NAFNET is designed for lightweight real-time deblurring of images.
15
 
16
  This is based on the implementation of NAFNet-DeBlur found [here](https://github.com/megvii-research/NAFNet.git).
17
+ 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).
18
 
19
  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.
20
 
 
27
 
28
  | Runtime | Precision | Chipset | SDK Versions | Download |
29
  |---|---|---|---|---|
30
+ | 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)
31
+ | 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)
32
+ | 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)
33
+ | 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)
34
+ | 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)
35
 
36
  For more device-specific assets and performance metrics, visit **[NAFNet-DeBlur on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/nafnet_deblur)**.
37
 
38
 
39
  ### Option 2: Export with Custom Configurations
40
 
41
+ 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:
42
  - Custom weights (e.g., fine-tuned checkpoints)
43
  - Custom input shapes
44
  - Target device and runtime configurations
45
 
46
  This option is ideal if you need to customize the model beyond the default configuration provided here.
47
 
48
+ 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.
49
 
50
  ## Model Details
51
 
 
60
  ## Performance Summary
61
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
62
  |---|---|---|---|---|---|---
63
+ | NAFNet-DeBlur | ONNX | float | Snapdragon® X2 Elite | 47.739 ms | 11 - 11 MB | NPU
64
+ | NAFNet-DeBlur | ONNX | float | Snapdragon® X Elite | 113.074 ms | 142 - 142 MB | NPU
65
+ | NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 81.438 ms | 0 - 1195 MB | NPU
66
+ | NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 183.315 ms | 10 - 862 MB | NPU
67
+ | NAFNet-DeBlur | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 112.679 ms | 0 - 161 MB | NPU
68
+ | NAFNet-DeBlur | ONNX | float | Qualcomm® QCS8450 | 183.315 ms | 10 - 862 MB | NPU
69
+ | NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 45.045 ms | 7 - 995 MB | NPU
70
+ | NAFNet-DeBlur | ONNX | float | Snapdragon® 8 Elite Mobile | 63.561 ms | 8 - 941 MB | NPU
71
+ | NAFNet-DeBlur | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 63.561 ms | 8 - 941 MB | NPU
72
+ | NAFNet-DeBlur | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 113.074 ms | 142 - 142 MB | NPU
73
+ | NAFNet-DeBlur | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 152.496 ms | 9 - 54 MB | NPU
74
+ | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® X2 Elite | 43.843 ms | 5 - 5 MB | NPU
75
+ | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® X Elite | 92.128 ms | 77 - 77 MB | NPU
76
+ | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 64.767 ms | 4 - 1084 MB | NPU
77
+ | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 89.707 ms | 0 - 84 MB | NPU
78
+ | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 54.509 ms | 4 - 967 MB | NPU
79
+ | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 105.292 ms | 3 - 7 MB | NPU
80
+ | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 133.706 ms | 6 - 1140 MB | NPU
81
+ | NAFNet-DeBlur | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 39.539 ms | 2 - 992 MB | NPU
82
+ | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 1326.727 ms | 0 - 1528 MB | NPU
83
+ | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 133.706 ms | 6 - 1140 MB | NPU
84
+ | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 54.509 ms | 4 - 967 MB | NPU
85
+ | NAFNet-DeBlur | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 92.128 ms | 77 - 77 MB | NPU
86
+ | NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® X2 Elite | 19.589 ms | 4 - 4 MB | NPU
87
+ | NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® X Elite | 46.725 ms | 74 - 74 MB | NPU
88
+ | NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 30.415 ms | 0 - 813 MB | NPU
89
+ | NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® 8 Gen 1 Mobile | 85.307 ms | 3 - 848 MB | NPU
90
+ | NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 223.392 ms | 1 - 5 MB | NPU
91
+ | NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 44.802 ms | 0 - 83 MB | NPU
92
+ | NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® QCS8450 | 85.307 ms | 3 - 848 MB | NPU
93
+ | NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® 8 Elite Mobile | 26.793 ms | 0 - 598 MB | NPU
94
+ | NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 59.482 ms | 2 - 835 MB | NPU
95
+ | NAFNet-DeBlur | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 18.968 ms | 0 - 748 MB | NPU
96
+ | NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 515.518 ms | 0 - 941 MB | NPU
97
+ | NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 59.482 ms | 2 - 835 MB | NPU
98
+ | NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 26.793 ms | 0 - 598 MB | NPU
99
+ | NAFNet-DeBlur | ONNX | w8a8 | Qualcomm® Dragonwing™ IQ-X7181 | 46.725 ms | 74 - 74 MB | NPU
100
+ | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® X2 Elite | 47.531 ms | 3 - 3 MB | NPU
101
+ | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® X Elite | 112.583 ms | 3 - 3 MB | NPU
102
+ | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 81.126 ms | 3 - 1202 MB | NPU
103
+ | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 185.051 ms | 2 - 839 MB | NPU
104
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® QCS8275 | 483.078 ms | 2 - 913 MB | NPU
105
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 109.221 ms | 3 - 6 MB | NPU
106
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA8775P | 139.95 ms | 1 - 917 MB | NPU
107
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA8650P | 139.95 ms | 1 - 917 MB | NPU
108
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA8255P | 139.95 ms | 1 - 917 MB | NPU
109
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® QCS8450 | 185.051 ms | 2 - 839 MB | NPU
110
+ | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 44.648 ms | 0 - 1003 MB | NPU
111
+ | NAFNet-DeBlur | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 63.005 ms | 0 - 948 MB | NPU
112
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA7255P | 483.078 ms | 2 - 913 MB | NPU
113
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® SA8295P | 162.305 ms | 0 - 632 MB | NPU
114
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 63.005 ms | 0 - 948 MB | NPU
115
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 112.583 ms | 3 - 3 MB | NPU
116
+ | NAFNet-DeBlur | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 152.687 ms | 3 - 10 MB | NPU
117
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 43.227 ms | 1 - 1 MB | NPU
118
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® X Elite | 91.189 ms | 1 - 1 MB | NPU
119
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 62.713 ms | 0 - 1032 MB | NPU
120
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® QCS8275 | 192.769 ms | 2 - 780 MB | NPU
121
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 88.626 ms | 2 - 5 MB | NPU
122
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® SA8775P | 86.069 ms | 2 - 780 MB | NPU
123
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® SA8650P | 86.069 ms | 2 - 780 MB | NPU
124
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® SA8255P | 86.069 ms | 2 - 780 MB | NPU
125
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 8 Elite Mobile | 52.175 ms | 1 - 934 MB | NPU
126
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 85.581 ms | 3 - 7 MB | NPU
127
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® SA7255P | 192.769 ms | 2 - 780 MB | NPU
128
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 136.397 ms | 1 - 1190 MB | NPU
129
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 39.251 ms | 1 - 1044 MB | NPU
130
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 1294.492 ms | 1 - 1486 MB | NPU
131
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 136.397 ms | 1 - 1190 MB | NPU
132
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 52.175 ms | 1 - 934 MB | NPU
133
+ | NAFNet-DeBlur | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 91.189 ms | 1 - 1 MB | NPU
134
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 21.311 ms | 1 - 1 MB | NPU
135
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® X Elite | 47.39 ms | 1 - 1 MB | NPU
136
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 30.401 ms | 1 - 786 MB | NPU
137
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® 8 Gen 1 Mobile | 85.808 ms | 1 - 812 MB | NPU
138
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 217.601 ms | 1 - 4 MB | NPU
139
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® QCS8275 | 91.145 ms | 1 - 605 MB | NPU
140
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 45.054 ms | 1 - 68 MB | NPU
141
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® SA8775P | 44.383 ms | 1 - 605 MB | NPU
142
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® SA8650P | 44.383 ms | 1 - 605 MB | NPU
143
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® SA8255P | 44.383 ms | 1 - 605 MB | NPU
144
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® QCS8450 | 85.808 ms | 1 - 812 MB | NPU
145
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® 8 Elite Mobile | 27.487 ms | 1 - 602 MB | NPU
146
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 56.27 ms | 1 - 827 MB | NPU
147
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 44.56 ms | 1 - 4 MB | NPU
148
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® SA7255P | 91.145 ms | 1 - 605 MB | NPU
149
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® SA8295P | 62.501 ms | 1 - 603 MB | NPU
150
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 17.199 ms | 1 - 695 MB | NPU
151
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 493.208 ms | 1 - 963 MB | NPU
152
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 56.27 ms | 1 - 827 MB | NPU
153
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 27.487 ms | 1 - 602 MB | NPU
154
+ | NAFNet-DeBlur | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-X7181 | 47.39 ms | 1 - 1 MB | NPU
155
+ | NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 80.988 ms | 3 - 1292 MB | NPU
156
+ | NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 210.907 ms | 4 - 1008 MB | NPU
157
+ | NAFNet-DeBlur | TFLITE | float | Qualcomm® QCS8275 | 483.041 ms | 3 - 1001 MB | NPU
158
+ | NAFNet-DeBlur | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 110.539 ms | 3 - 7 MB | NPU
159
+ | NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8775P | 139.99 ms | 3 - 1005 MB | NPU
160
+ | NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8650P | 139.99 ms | 3 - 1005 MB | NPU
161
+ | NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8255P | 139.99 ms | 3 - 1005 MB | NPU
162
+ | NAFNet-DeBlur | TFLITE | float | Qualcomm® QCS8450 | 210.907 ms | 4 - 1008 MB | NPU
163
+ | NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 44.656 ms | 3 - 1090 MB | NPU
164
+ | NAFNet-DeBlur | TFLITE | float | Snapdragon® 8 Elite Mobile | 62.301 ms | 2 - 1032 MB | NPU
165
+ | NAFNet-DeBlur | TFLITE | float | Qualcomm® SA7255P | 483.041 ms | 3 - 1001 MB | NPU
166
+ | NAFNet-DeBlur | TFLITE | float | Qualcomm® SA8295P | 162.41 ms | 3 - 738 MB | NPU
167
+ | NAFNet-DeBlur | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 62.301 ms | 2 - 1032 MB | NPU
168
+ | NAFNet-DeBlur | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 152.356 ms | 1 - 150 MB | NPU
169
+ | NAFNet-DeBlur | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 60.273 ms | 1 - 1099 MB | NPU
170
+ | NAFNet-DeBlur | TFLITE | w8a8 | Snapdragon® 8 Gen 1 Mobile | 123.385 ms | 1 - 1192 MB | NPU
171
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® QCS8275 | 169.472 ms | 1 - 868 MB | NPU
172
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 84.724 ms | 1 - 5 MB | NPU
173
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® SA8775P | 84.58 ms | 1 - 868 MB | NPU
174
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® SA8650P | 84.58 ms | 1 - 868 MB | NPU
175
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® SA8255P | 84.58 ms | 1 - 868 MB | NPU
176
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® QCS8450 | 123.385 ms | 1 - 1192 MB | NPU
177
+ | NAFNet-DeBlur | TFLITE | w8a8 | Snapdragon® 8 Elite Mobile | 40.239 ms | 1 - 703 MB | NPU
178
+ | NAFNet-DeBlur | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 885.229 ms | 138 - 542 MB | NPU
179
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 109.926 ms | 1 - 81 MB | NPU
180
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® SA7255P | 169.472 ms | 1 - 868 MB | NPU
181
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® SA8295P | 116.821 ms | 1 - 866 MB | NPU
182
+ | NAFNet-DeBlur | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 31.605 ms | 1 - 1209 MB | NPU
183
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 885.229 ms | 138 - 542 MB | NPU
184
+ | NAFNet-DeBlur | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 40.239 ms | 1 - 703 MB | NPU
185
 
186
  ## License
187
  * The license for the original implementation of NAFNet-DeBlur can be found
release_assets.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "version": "0.57.3",
3
  "precisions": {
4
  "w8a16": {
5
  "universal_assets": {
@@ -7,14 +7,14 @@
7
  "tool_versions": {
8
  "qairt": "2.45.0.260326154327"
9
  },
10
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_deblur/releases/v0.57.3/nafnet_deblur-qnn_dlc-w8a16.zip"
11
  },
12
  "onnx": {
13
  "tool_versions": {
14
  "qairt": "2.45.0.260326154327",
15
  "onnx_runtime": "1.25.0"
16
  },
17
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_deblur/releases/v0.57.3/nafnet_deblur-onnx-w8a16.zip"
18
  }
19
  }
20
  },
@@ -25,20 +25,20 @@
25
  "qairt": "2.45.0.260326154327",
26
  "litert": "1.4.4"
27
  },
28
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_deblur/releases/v0.57.3/nafnet_deblur-tflite-float.zip"
29
  },
30
  "qnn_dlc": {
31
  "tool_versions": {
32
  "qairt": "2.45.0.260326154327"
33
  },
34
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_deblur/releases/v0.57.3/nafnet_deblur-qnn_dlc-float.zip"
35
  },
36
  "onnx": {
37
  "tool_versions": {
38
  "qairt": "2.45.0.260326154327",
39
  "onnx_runtime": "1.25.0"
40
  },
41
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/nafnet_deblur/releases/v0.57.3/nafnet_deblur-onnx-float.zip"
42
  }
43
  }
44
  }
 
1
  {
2
+ "version": "0.58.0",
3
  "precisions": {
4
  "w8a16": {
5
  "universal_assets": {
 
7
  "tool_versions": {
8
  "qairt": "2.45.0.260326154327"
9
  },
10
+ "download_url": "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"
11
  },
12
  "onnx": {
13
  "tool_versions": {
14
  "qairt": "2.45.0.260326154327",
15
  "onnx_runtime": "1.25.0"
16
  },
17
+ "download_url": "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"
18
  }
19
  }
20
  },
 
25
  "qairt": "2.45.0.260326154327",
26
  "litert": "1.4.4"
27
  },
28
+ "download_url": "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"
29
  },
30
  "qnn_dlc": {
31
  "tool_versions": {
32
  "qairt": "2.45.0.260326154327"
33
  },
34
+ "download_url": "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"
35
  },
36
  "onnx": {
37
  "tool_versions": {
38
  "qairt": "2.45.0.260326154327",
39
  "onnx_runtime": "1.25.0"
40
  },
41
+ "download_url": "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"
42
  }
43
  }
44
  }