Hugging Face's logo

qualcomm
/
QuickSRNetMedium

Image-to-Image
PyTorch
android
Model card Files
xet
 Community
QuickSRNetMedium
File size: 10,418 Bytes
d73fff9
 
 
 
 
cbc1c2e
d73fff9
 
 
5d2cbcb
d73fff9
ba48e8f
b997c08
d73fff9
 
ba48e8f
4874fbd
ba48e8f
 
 
 
 
 
 
 
 
 
 
 
4874fbd
 
 
 
 
 
ba48e8f
 
 
 
 
 
4874fbd
ba48e8f
 
 
 
 
 
4874fbd
ba48e8f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4874fbd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ba48e8f
4874fbd
 
 
 
 
 
 
 
 
 
 
 
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
---
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/tree/v0.49.1/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 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.49.1/quicksrnetmedium-onnx-float.zip)
| ONNX | w8a8 | Universal | QAIRT 2.42, ONNX Runtime 1.24.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.49.1/quicksrnetmedium-onnx-w8a8.zip)
| QNN_DLC | float | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.49.1/quicksrnetmedium-qnn_dlc-float.zip)
| QNN_DLC | w8a8 | Universal | QAIRT 2.43 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.49.1/quicksrnetmedium-qnn_dlc-w8a8.zip)
| TFLITE | float | Universal | QAIRT 2.43, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.49.1/quicksrnetmedium-tflite-float.zip)
| TFLITE | w8a8 | Universal | QAIRT 2.43, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/quicksrnetmedium/releases/v0.49.1/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/tree/v0.49.1/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/tree/v0.49.1/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® 8 Elite Gen 5 Mobile | 0.448 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® X2 Elite | 0.463 ms | 6 - 6 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® X Elite | 1.095 ms | 9 - 9 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.661 ms | 0 - 28 MB | NPU
| QuickSRNetMedium | ONNX | float | Qualcomm® QCS8550 (Proxy) | 0.972 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | ONNX | float | Qualcomm® QCS9075 | 1.329 ms | 6 - 9 MB | NPU
| QuickSRNetMedium | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.548 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.23 ms | 0 - 23 MB | CPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® X2 Elite | 0.221 ms | 3 - 3 MB | CPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® X Elite | 0.619 ms | 3 - 3 MB | CPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.343 ms | 0 - 29 MB | CPU
| QuickSRNetMedium | ONNX | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.516 ms | 0 - 16 MB | CPU
| QuickSRNetMedium | ONNX | w8a8 | Qualcomm® QCS9075 | 0.687 ms | 0 - 3 MB | CPU
| QuickSRNetMedium | ONNX | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.273 ms | 0 - 18 MB | CPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.326 ms | 0 - 24 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® X2 Elite | 0.468 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® X Elite | 0.894 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 0.495 ms | 0 - 29 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 2.549 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 0.84 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® SA8775P | 1.208 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® QCS9075 | 1.192 ms | 2 - 7 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 1.159 ms | 0 - 31 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® SA7255P | 2.549 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Qualcomm® SA8295P | 1.55 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.385 ms | 0 - 24 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.154 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 0.238 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® X Elite | 0.47 ms | 0 - 0 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.221 ms | 0 - 26 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS6490 | 1.338 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS8275 (Proxy) | 0.839 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.343 ms | 0 - 1 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA8775P | 0.537 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS9075 | 0.494 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCM6690 | 1.564 ms | 0 - 17 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® QCS8450 (Proxy) | 0.534 ms | 0 - 27 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA7255P | 0.839 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Qualcomm® SA8295P | 0.738 ms | 0 - 16 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.182 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.371 ms | 0 - 17 MB | NPU
| QuickSRNetMedium | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.412 ms | 0 - 24 MB | NPU
| QuickSRNetMedium | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 0.673 ms | 0 - 30 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 3.089 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 1.078 ms | 0 - 2 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® SA8775P | 1.551 ms | 0 - 22 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® QCS9075 | 1.412 ms | 1 - 6 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 1.454 ms | 0 - 30 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® SA7255P | 3.089 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | TFLITE | float | Qualcomm® SA8295P | 1.795 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.5 ms | 0 - 25 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.176 ms | 0 - 21 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.289 ms | 0 - 27 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS6490 | 1.166 ms | 0 - 3 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS8275 (Proxy) | 0.958 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.438 ms | 0 - 1 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA8775P | 0.637 ms | 0 - 20 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS9075 | 0.568 ms | 0 - 3 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCM6690 | 1.943 ms | 0 - 17 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® QCS8450 (Proxy) | 0.506 ms | 0 - 28 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA7255P | 0.958 ms | 0 - 19 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Qualcomm® SA8295P | 0.854 ms | 0 - 16 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.233 ms | 0 - 18 MB | NPU
| QuickSRNetMedium | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.461 ms | 0 - 18 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).