Hugging Face's logo

qualcomm
/
XLSR

Image-to-Image
PyTorch
LiteRT
android
Model card Files Community
1
XLSR
File size: 18,040 Bytes 
96531c5
 
 
 
 
5130323
96531c5
 
 
fbc84ae
96531c5
 
 
 
85bce41
96531c5
 
05aa505
85bce41
 
96531c5
 
 
 
 
 
 
6507118
96531c5
1277a33
01cde09
0ec5bc8
 
 
1277a33
6507118
8f9afb6
450b98f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
38d5887
450b98f
 
38d5887
450b98f
 
38d5887
450b98f
38d5887
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0ec5bc8
38d5887
 
 
 
 
 
 
 
 
1277a33
96531c5
 
 
 
 
 
8a76801
96531c5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
77a8192
8f9afb6
 
 
6507118
 
450b98f
f2e13b5
6507118
 
77a8192
1277a33
 
96531c5
 
1277a33
96531c5
 
 
 
 
 
 
 
 
 
 
 
3f73005
96531c5
 
3f73005
96531c5
 
8a76801
96531c5
3f73005
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96531c5
 
 
 
 
 
 
 
 
 
 
 
4c3b0cf
 
 
3bdbf30
96531c5
 
 
 
 
 
 
 
 
4c3b0cf
 
 
 
3bdbf30
96531c5
 
 
 
 
 
fbc84ae
96531c5
 
 
3bdbf30
 
 
 
 
f2e13b5
3bdbf30
 
 
 
 
f2e13b5
3bdbf30
 
1277a33
96531c5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8f9afb6
96531c5
8a76801
 
8f9afb6
 
 
96531c5
 
 
 
 
8f9afb6
 
96531c5
f8cd1cc
96531c5
 
 
 
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
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
 
---
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/xlsr/web-assets/model_demo.png)

# XLSR: Optimized for Mobile Deployment
## Upscale images in real time


XLSR is designed for lightweight real-time upscaling of images.

This model is an implementation of XLSR found [here](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/xlsr).


This repository provides scripts to run XLSR on Qualcomm® devices.
More details on model performance across various devices, can be found
[here](https://aihub.qualcomm.com/models/xlsr).


### Model Details

- **Model Type:** Model_use_case.super_resolution
- **Model Stats:**
  - Model checkpoint: xlsr_3x_checkpoint
  - Input resolution: 128x128
  - Number of parameters: 28.0K
  - Model size (float): 115 KB
  - Model size (w8a8): 45.6 KB

| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
|---|---|---|---|---|---|---|---|---|
| XLSR | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 8.429 ms | 6 - 18 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 2.31 ms | 0 - 12 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 3.693 ms | 6 - 34 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1.136 ms | 0 - 25 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 2.9 ms | 0 - 6 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 0.865 ms | 0 - 6 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 4.152 ms | 6 - 20 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.228 ms | 0 - 14 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 8.429 ms | 6 - 18 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 2.31 ms | 0 - 12 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 2.999 ms | 0 - 9 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 0.81 ms | 0 - 6 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 4.541 ms | 6 - 21 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 1.312 ms | 0 - 21 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 2.849 ms | 0 - 8 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 0.837 ms | 0 - 6 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 4.152 ms | 6 - 20 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.228 ms | 0 - 14 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 2.878 ms | 0 - 4 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 0.825 ms | 0 - 5 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 1.36 ms | 0 - 7 MB | NPU | [XLSR.onnx](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.onnx) |
| XLSR | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.95 ms | 0 - 25 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.474 ms | 0 - 24 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 0.833 ms | 0 - 24 MB | NPU | [XLSR.onnx](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.onnx) |
| XLSR | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 2.169 ms | 0 - 15 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.tflite) |
| XLSR | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 0.528 ms | 0 - 18 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 0.8 ms | 1 - 15 MB | NPU | [XLSR.onnx](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.onnx) |
| XLSR | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 0.919 ms | 0 - 0 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.dlc) |
| XLSR | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.417 ms | 8 - 8 MB | NPU | [XLSR.onnx](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR.onnx) |
| XLSR | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 2.451 ms | 1 - 11 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 0.9 ms | 0 - 11 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 1.991 ms | 0 - 25 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 0.524 ms | 0 - 24 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.035 ms | 0 - 16 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 0.392 ms | 0 - 11 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.755 ms | 0 - 11 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 0.603 ms | 0 - 13 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | TFLITE | 2.104 ms | 0 - 15 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 1.105 ms | 0 - 15 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | TFLITE | 15.922 ms | 4 - 14 MB | GPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 2.451 ms | 1 - 11 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 0.9 ms | 0 - 11 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 1.13 ms | 1 - 17 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 0.388 ms | 0 - 3 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 1.73 ms | 0 - 19 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 0.744 ms | 0 - 14 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 1.045 ms | 0 - 17 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 0.414 ms | 0 - 11 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 1.755 ms | 0 - 11 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 0.603 ms | 0 - 13 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 1.038 ms | 0 - 17 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 0.406 ms | 0 - 10 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 1.243 ms | 0 - 16 MB | NPU | [XLSR.onnx](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.onnx) |
| XLSR | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 0.846 ms | 0 - 24 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.24 ms | 0 - 25 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 0.838 ms | 0 - 20 MB | NPU | [XLSR.onnx](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.onnx) |
| XLSR | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 1.321 ms | 0 - 17 MB | NPU | [XLSR.tflite](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.tflite) |
| XLSR | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 0.245 ms | 0 - 19 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 0.81 ms | 3 - 22 MB | NPU | [XLSR.onnx](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.onnx) |
| XLSR | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 0.526 ms | 5 - 5 MB | NPU | [XLSR.dlc](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.dlc) |
| XLSR | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.322 ms | 9 - 9 MB | NPU | [XLSR.onnx](https://huggingface.co/qualcomm/XLSR/blob/main/XLSR_w8a8.onnx) |




## Installation


Install the package via pip:
```bash
pip install qai-hub-models
```


## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device

Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your
Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.

With this API token, you can configure your client to run models on the cloud
hosted devices.
```bash
qai-hub configure --api_token API_TOKEN
```
Navigate to [docs](https://app.aihub.qualcomm.com/docs/) for more information.



## Demo off target

The package contains a simple end-to-end demo that downloads pre-trained
weights and runs this model on a sample input.

```bash
python -m qai_hub_models.models.xlsr.demo
```

The above demo runs a reference implementation of pre-processing, model
inference, and post processing.

**NOTE**: If you want running in a Jupyter Notebook or Google Colab like
environment, please add the following to your cell (instead of the above).
```
%run -m qai_hub_models.models.xlsr.demo
```


### Run model on a cloud-hosted device

In addition to the demo, you can also run the model on a cloud-hosted Qualcomm®
device. This script does the following:
* Performance check on-device on a cloud-hosted device
* Downloads compiled assets that can be deployed on-device for Android.
* Accuracy check between PyTorch and on-device outputs.

```bash
python -m qai_hub_models.models.xlsr.export
```
```
Profiling Results
------------------------------------------------------------
XLSR
Device                          : cs_8275 (ANDROID 14)                
Runtime                         : TFLITE                              
Estimated inference time (ms)   : 8.4                                 
Estimated peak memory usage (MB): [6, 18]                             
Total # Ops                     : 17                                  
Compute Unit(s)                 : npu (14 ops) gpu (0 ops) cpu (3 ops)
```


## How does this work?

This [export script](https://aihub.qualcomm.com/models/xlsr/qai_hub_models/models/XLSR/export.py)
leverages [Qualcomm® AI Hub](https://aihub.qualcomm.com/) to optimize, validate, and deploy this model
on-device. Lets go through each step below in detail:

Step 1: **Compile model for on-device deployment**

To compile a PyTorch model for on-device deployment, we first trace the model
in memory using the `jit.trace` and then call the `submit_compile_job` API.

```python
import torch

import qai_hub as hub
from qai_hub_models.models.xlsr import Model

# Load the model
torch_model = Model.from_pretrained()

# Device
device = hub.Device("Samsung Galaxy S24")

# Trace model
input_shape = torch_model.get_input_spec()
sample_inputs = torch_model.sample_inputs()

pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])

# Compile model on a specific device
compile_job = hub.submit_compile_job(
    model=pt_model,
    device=device,
    input_specs=torch_model.get_input_spec(),
)

# Get target model to run on-device
target_model = compile_job.get_target_model()

```


Step 2: **Performance profiling on cloud-hosted device**

After compiling models from step 1. Models can be profiled model on-device using the
`target_model`. Note that this scripts runs the model on a device automatically
provisioned in the cloud.  Once the job is submitted, you can navigate to a
provided job URL to view a variety of on-device performance metrics.
```python
profile_job = hub.submit_profile_job(
    model=target_model,
    device=device,
)
        
```

Step 3: **Verify on-device accuracy**

To verify the accuracy of the model on-device, you can run on-device inference
on sample input data on the same cloud hosted device.
```python
input_data = torch_model.sample_inputs()
inference_job = hub.submit_inference_job(
    model=target_model,
    device=device,
    inputs=input_data,
)
    on_device_output = inference_job.download_output_data()

```
With the output of the model, you can compute like PSNR, relative errors or
spot check the output with expected output.

**Note**: This on-device profiling and inference requires access to Qualcomm®
AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).



## Run demo on a cloud-hosted device

You can also run the demo on-device.

```bash
python -m qai_hub_models.models.xlsr.demo --eval-mode on-device
```

**NOTE**: If you want running in a Jupyter Notebook or Google Colab like
environment, please add the following to your cell (instead of the above).
```
%run -m qai_hub_models.models.xlsr.demo -- --eval-mode on-device
```


## Deploying compiled model to Android


The models can be deployed using multiple runtimes:
- TensorFlow Lite (`.tflite` export): [This
  tutorial](https://www.tensorflow.org/lite/android/quickstart) provides a
  guide to deploy the .tflite model in an Android application.


- QNN (`.so` export ): This [sample
  app](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html)
provides instructions on how to use the `.so` shared library  in an Android application.


## View on Qualcomm® AI Hub
Get more details on XLSR's performance across various devices [here](https://aihub.qualcomm.com/models/xlsr).
Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)


## License
* The license for the original implementation of XLSR can be found
  [here](https://github.com/quic/aimet-model-zoo/blob/develop/LICENSE.pdf).
* The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf)



## References
* [Extremely Lightweight Quantization Robust Real-Time Single-Image Super Resolution for Mobile Devices](https://arxiv.org/abs/2105.10288)
* [Source Model Implementation](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/xlsr)



## 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:[email protected]).