Hugging Face's logo

qualcomm
/
QuickSRNetMedium

Image-to-Image
PyTorch
android
Model card Files
xet
 Community
QuickSRNetMedium
File size: 12,865 Bytes
d73fff9
 
 
 
 
cbc1c2e
d73fff9
 
 
5d2cbcb
d73fff9
ba48e8f
b997c08
d73fff9
 
ba48e8f
ddc22c1
ba48e8f
 
 
 
 
 
 
 
 
 
 
 
ddc22c1
 
 
 
 
 
ba48e8f
 
 
 
 
 
ddc22c1
ba48e8f
 
 
 
 
 
ddc22c1
ba48e8f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ddc22c1
 
 
 
 
 
 
 
 
 
 
d05054a
ddc22c1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d3d01c5
d73fff9
56de2b6
4874fbd
d3d01c5
d73fff9
 
 
 
 
e241025
d73fff9
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
---
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/quicksrnetmedium/web-assets/model_demo.png)

# QuickSRNetMedium: Optimized for Qualcomm Devices

QuickSRNet Medium is designed for upscaling images on mobile platforms to sharpen in real-time.

This is based on the implementation of QuickSRNetMedium found [here](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/quicksrnet).
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/quicksrnetmedium) 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/quicksrnetmedium/releases/v0.59.0/quicksrnetmedium-onnx-float.zip)
| ONNX | w8a8 | Universal | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.59.0/quicksrnetmedium-onnx-w8a8.zip)
| QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.59.0/quicksrnetmedium-qnn_dlc-float.zip)
| QNN_DLC | w8a8 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.59.0/quicksrnetmedium-qnn_dlc-w8a8.zip)
| TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.59.0/quicksrnetmedium-tflite-float.zip)
| TFLITE | w8a8 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.59.0/quicksrnetmedium-tflite-w8a8.zip)

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


### 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/quicksrnetmedium) 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 [QuickSRNetMedium on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.59.0/src/qai_hub_models/models/quicksrnetmedium) for usage instructions.

## Model Details

**Model Type:** Model_use_case.super_resolution

**Model Stats:**
- Model checkpoint: quicksrnet_medium_3x_checkpoint
- Input resolution: 128x128
- Number of parameters: 61.0K
- Model size (float): 243 KB
- Model size (w8a8): 73.9 KB

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| QuickSRNetMedium | ONNX | float | Snapdragon® X2 Elite | 0.429 ms | 8 - 8 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® X Elite | 0.928 ms | 8 - 8 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.6 ms | 0 - 29 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 1.368 ms | 0 - 35 MB | NPU
| QuickSRNetMedium | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.909 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | ONNX | float | Qualcomm® QCS8450 | 1.368 ms | 0 - 35 MB | NPU
| QuickSRNetMedium | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 1.211 ms | 6 - 9 MB | NPU
| QuickSRNetMedium | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.928 ms | 8 - 8 MB | NPU
| QuickSRNetMedium | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 0.499 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® 8 Elite Mobile | 0.499 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.431 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® X2 Elite | 0.204 ms | 3 - 3 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® X Elite | 0.501 ms | 3 - 3 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.34 ms | 0 - 30 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.62 ms | 0 - 35 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 1.812 ms | 0 - 3 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.503 ms | 0 - 24 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Qualcomm® QCS8450 | 0.62 ms | 0 - 35 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 0.652 ms | 0 - 3 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Qualcomm® Dragonwing™ IQ-X7181 | 0.501 ms | 3 - 3 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 0.245 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® 8 Elite Mobile | 0.245 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.221 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® X2 Elite | 0.468 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® X Elite | 0.895 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 0.49 ms | 0 - 30 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 1.218 ms | 0 - 35 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.799 ms | 0 - 28 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® SA8775P | 1.185 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® SA8650P | 1.185 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® SA8255P | 1.185 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® QCS8450 | 1.218 ms | 0 - 35 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 1.077 ms | 0 - 5 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.895 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 0.385 ms | 0 - 25 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® SA8295P | 1.566 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 0.385 ms | 0 - 25 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.37 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 0.244 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® X Elite | 0.447 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.229 ms | 0 - 29 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.445 ms | 0 - 34 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 1.408 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS8275 | 0.9 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.362 ms | 0 - 1 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA8775P | 0.552 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA8650P | 0.552 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA8255P | 0.552 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS8450 | 0.445 ms | 0 - 34 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 0.477 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-X7181 | 0.447 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 1.741 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 0.392 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 0.199 ms | 0 - 25 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA7255P | 0.9 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA8295P | 0.733 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 8 Elite Mobile | 0.199 ms | 0 - 25 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.171 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.392 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 0.666 ms | 0 - 30 MB | NPU
| QuickSRNetMedium | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 1.465 ms | 0 - 36 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 1.075 ms | 0 - 4 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® SA8775P | 1.556 ms | 0 - 23 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® SA8650P | 1.556 ms | 0 - 23 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® SA8255P | 1.556 ms | 0 - 23 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® QCS8450 | 1.465 ms | 0 - 36 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 1.43 ms | 3 - 8 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 0.515 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® SA8295P | 1.826 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | TFLITE | float | Snapdragon® 8 Elite Mobile | 0.515 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.431 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.293 ms | 0 - 29 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.521 ms | 0 - 34 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 1.247 ms | 0 - 3 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS8275 | 0.994 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.453 ms | 0 - 16 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA8775P | 0.64 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA8650P | 0.64 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA8255P | 0.64 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS8450 | 0.521 ms | 0 - 34 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 0.585 ms | 0 - 3 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 1.928 ms | 0 - 17 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 0.481 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 0.233 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA7255P | 0.994 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA8295P | 0.874 ms | 0 - 17 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 8 Elite Mobile | 0.233 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.195 ms | 0 - 23 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.481 ms | 0 - 19 MB | NPU

## License
* The license for the original implementation of QuickSRNetMedium can be found
  [here](https://github.com/quic/aimet-model-zoo/blob/develop/LICENSE.md).

## References
* [QuickSRNet: Plain Single-Image Super-Resolution Architecture for Faster Inference on Mobile Platforms](https://arxiv.org/abs/2303.04336)
* [Source Model Implementation](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/quicksrnet)

## 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).