Hugging Face's logo

qualcomm
/
VIT

Image Classification
PyTorch
LiteRT
backbone
android
Model card Files
xet
Community
VIT
File size: 21,080 Bytes 
b7b7202
 
1d15e1e
b7b7202
 
 
878bbb1
b7b7202
 
 
cb4fccc
b7b7202
 
 
 
fff7db8
b7b7202
 
6c96959
fff7db8
 
b7b7202
 
 
 
 
 
 
1d15e1e
b7b7202
 
 
 
cd90639
 
 
b7b7202
1d15e1e
e55e123
d8f9804
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
cd90639
d8f9804
 
cd90639
d8f9804
 
cd90639
d8f9804
cd90639
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b7b7202
488f76a
 
b7b7202
 
 
 
36c6a95
b7b7202
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e55e123
 
 
 
1d15e1e
 
d8f9804
ac709b4
1d15e1e
 
e55e123
5265d22
 
b7b7202
 
488f76a
b7b7202
 
 
 
 
 
 
 
 
 
 
 
9870567
b7b7202
 
9870567
b7b7202
 
36c6a95
b7b7202
9870567
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b7b7202
 
 
 
 
 
 
 
 
 
 
 
80ece63
 
 
3cc962d
b7b7202
 
 
 
 
 
 
 
 
80ece63
 
 
 
3cc962d
b7b7202
 
 
 
 
 
cb4fccc
b7b7202
 
488f76a
b7b7202
 
 
 
 
ac709b4
b7b7202
 
 
 
 
ac709b4
b7b7202
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e55e123
b7b7202
36c6a95
 
e55e123
 
 
b7b7202
 
 
 
 
e55e123
 
b7b7202
221f340
b7b7202
 
 
 
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
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
 
---
library_name: pytorch
license: other
tags:
- backbone
- android
pipeline_tag: image-classification

---

![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/vit/web-assets/model_demo.png)

# VIT: Optimized for Mobile Deployment
## Imagenet classifier and general purpose backbone


VIT is a machine learning model that can classify images from the Imagenet dataset. It can also be used as a backbone in building more complex models for specific use cases.

This model is an implementation of VIT found [here](https://github.com/pytorch/vision/blob/main/torchvision/models/vision_transformer.py).


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


### Model Details

- **Model Type:** Model_use_case.image_classification
- **Model Stats:**
  - Model checkpoint: Imagenet
  - Input resolution: 224x224
  - Number of parameters: 86.6M
  - Model size (float): 330 MB
  - Model size (w8a16): 86.2 MB
  - Model size (w8a8): 83.2 MB

| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
|---|---|---|---|---|---|---|---|---|
| VIT | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 43.891 ms | 0 - 315 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 45.176 ms | 0 - 324 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 17.891 ms | 0 - 321 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 21.31 ms | 0 - 316 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 12.696 ms | 0 - 28 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 13.809 ms | 0 - 32 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 16.043 ms | 0 - 315 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 16.876 ms | 1 - 324 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 43.891 ms | 0 - 315 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 45.176 ms | 0 - 324 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 13.166 ms | 0 - 24 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 13.828 ms | 0 - 31 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 20.073 ms | 0 - 307 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 19.819 ms | 1 - 327 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 12.733 ms | 0 - 26 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 13.893 ms | 0 - 30 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 16.043 ms | 0 - 315 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 16.876 ms | 1 - 324 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 12.72 ms | 0 - 13 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 13.859 ms | 0 - 28 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 13.54 ms | 0 - 399 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT.onnx) |
| VIT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 8.994 ms | 0 - 319 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 9.587 ms | 38 - 370 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 9.308 ms | 10 - 344 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT.onnx) |
| VIT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 8.262 ms | 0 - 319 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT.tflite) |
| VIT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 8.024 ms | 1 - 314 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 7.829 ms | 1 - 321 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT.onnx) |
| VIT | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 16.174 ms | 1116 - 1116 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT.dlc) |
| VIT | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 14.919 ms | 171 - 171 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT.onnx) |
| VIT | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 64.963 ms | 0 - 189 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 51.313 ms | 0 - 210 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 25.858 ms | 0 - 48 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 22.689 ms | 0 - 189 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 201.78 ms | 0 - 1634 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 64.963 ms | 0 - 189 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 25.98 ms | 0 - 48 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 37.12 ms | 0 - 212 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 26.02 ms | 0 - 47 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 22.689 ms | 0 - 189 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 26.113 ms | 0 - 48 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 150.726 ms | 578 - 827 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.onnx) |
| VIT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 19.328 ms | 0 - 195 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 125.044 ms | 673 - 835 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.onnx) |
| VIT | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 16.013 ms | 0 - 186 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 106.848 ms | 662 - 793 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.onnx) |
| VIT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 27.231 ms | 379 - 379 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.dlc) |
| VIT | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 180.374 ms | 924 - 924 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a16.onnx) |
| VIT | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 24.381 ms | 0 - 48 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 30.273 ms | 0 - 164 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 12.796 ms | 0 - 57 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 16.349 ms | 0 - 227 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 12.099 ms | 3 - 117 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 10.505 ms | 0 - 27 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 12.516 ms | 0 - 49 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 9.592 ms | 0 - 164 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | TFLITE | 80.928 ms | 2 - 44 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 79.12 ms | 0 - 406 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 24.381 ms | 0 - 48 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 30.273 ms | 0 - 164 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 12.215 ms | 0 - 101 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 10.575 ms | 0 - 26 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 14.655 ms | 0 - 50 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 16.438 ms | 0 - 169 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 12.128 ms | 0 - 93 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 10.53 ms | 0 - 28 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 12.516 ms | 0 - 49 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 9.592 ms | 0 - 164 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 12.213 ms | 0 - 102 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 10.549 ms | 0 - 27 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 28.41 ms | 0 - 123 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.onnx) |
| VIT | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 8.69 ms | 0 - 53 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 6.985 ms | 0 - 167 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 20.049 ms | 0 - 278 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.onnx) |
| VIT | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 5.952 ms | 0 - 55 MB | NPU | [VIT.tflite](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.tflite) |
| VIT | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 6.077 ms | 0 - 163 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 17.14 ms | 0 - 254 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.onnx) |
| VIT | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 11.728 ms | 425 - 425 MB | NPU | [VIT.dlc](https://huggingface.co/qualcomm/VIT/blob/main/VIT_w8a8.dlc) |
| VIT | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 34.308 ms | 88 - 88 MB | NPU | [VIT.onnx](https://huggingface.co/qualcomm/VIT/blob/main/VIT_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.vit.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.vit.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.vit.export
```
```
Profiling Results
------------------------------------------------------------
VIT
Device                          : cs_8275 (ANDROID 14)                  
Runtime                         : TFLITE                                
Estimated inference time (ms)   : 43.9                                  
Estimated peak memory usage (MB): [0, 315]                              
Total # Ops                     : 1579                                  
Compute Unit(s)                 : npu (1579 ops) gpu (0 ops) cpu (0 ops)
```


## How does this work?

This [export script](https://aihub.qualcomm.com/models/vit/qai_hub_models/models/VIT/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.vit 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.vit.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.vit.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 VIT's performance across various devices [here](https://aihub.qualcomm.com/models/vit).
Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)


## License
* The license for the original implementation of VIT can be found
  [here](https://github.com/pytorch/vision/blob/main/LICENSE).
* 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
* [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929)
* [Source Model Implementation](https://github.com/pytorch/vision/blob/main/torchvision/models/vision_transformer.py)



## 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]).