Object Detection
PyTorch
real_time
android
File size: 14,156 Bytes
08e0b16
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
875be77
08e0b16
 
 
 
 
875be77
08e0b16
 
 
 
875be77
08e0b16
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
875be77
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
547132c
 
875be77
 
 
 
 
 
 
 
 
547132c
 
875be77
 
 
 
 
 
 
 
547132c
 
875be77
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
547132c
 
875be77
 
 
 
 
 
 
 
 
547132c
 
875be77
 
 
 
 
 
 
 
547132c
 
875be77
 
 
 
08e0b16
 
 
 
 
 
 
 
 
 
 
 
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
---
library_name: pytorch
license: gpl-3.0
tags:
- real_time
- android
pipeline_tag: object-detection

---

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

# WeDetect: Optimized for Qualcomm Devices

WeDetect is a text-conditioned object detection model that uses a prompt-then-detect strategy. Text class names are reparameterized into the model weights before export, producing a fixed-vocabulary detector that runs efficiently on-device. The model uses LTRB (Left-Top-Right-Bottom) regression with multi-scale feature maps at strides 8, 16, and 32.

This is based on the implementation of WeDetect found [here](https://github.com/WeChatCV/WeDetect).
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/v0.63.0/src/qai_hub_models/models/wedetect) 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
Due to licensing restrictions, we cannot distribute pre-exported model assets for this model.
Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.63.0/src/qai_hub_models/models/wedetect) 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

See our repository for [WeDetect on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.63.0/src/qai_hub_models/models/wedetect) for usage instructions.


## Model Details

**Model Type:** Model_use_case.object_detection

**Model Stats:**
- Input resolution: 640x640
- Model checkpoint: wedetect_tiny
- Model size (detector) (float): 143 MB
- Model size (detector) (mixed_with_float): 35.9 MB
- Model size (text_encoder) (float): 592 B
- Model size (text_encoder) (mixed_with_float): 13.3 KB
- Number of parameters (detector): 37.3M
- Number of parameters (text_encoder): 3.00

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| detector | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 9.235 ms | 1 - 244 MB | NPU
| detector | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 11.53 ms | 1 - 240 MB | NPU
| detector | ONNX | float | Snapdragon® X2 Elite | 11.417 ms | 5 - 5 MB | NPU
| detector | ONNX | float | Snapdragon® X Elite | 21.053 ms | 71 - 71 MB | NPU
| detector | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 15.889 ms | 0 - 319 MB | NPU
| detector | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 76.841 ms | 1 - 318 MB | NPU
| detector | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 34.96 ms | 5 - 14 MB | NPU
| detector | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 20.334 ms | 0 - 77 MB | NPU
| detector | ONNX | float | Qualcomm® QCS8450 | 76.841 ms | 1 - 318 MB | NPU
| detector | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 30.936 ms | 5 - 13 MB | NPU
| detector | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 21.053 ms | 71 - 71 MB | NPU
| detector | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 11.53 ms | 1 - 240 MB | NPU
| detector | ONNX | mixed_with_float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 7.682 ms | 4 - 283 MB | NPU
| detector | ONNX | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 9.35 ms | 7 - 281 MB | NPU
| detector | ONNX | mixed_with_float | Snapdragon® X2 Elite | 7.438 ms | 9 - 9 MB | NPU
| detector | ONNX | mixed_with_float | Snapdragon® X Elite | 16.328 ms | 34 - 34 MB | NPU
| detector | ONNX | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 11.155 ms | 12 - 367 MB | NPU
| detector | ONNX | mixed_with_float | Snapdragon® 8 Gen 1 Mobile | 24.199 ms | 11 - 367 MB | NPU
| detector | ONNX | mixed_with_float | Qualcomm® Dragonwing™ IQ-8275 | 15.51 ms | 6 - 15 MB | NPU
| detector | ONNX | mixed_with_float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 15.839 ms | 0 - 38 MB | NPU
| detector | ONNX | mixed_with_float | Qualcomm® QCS8450 | 24.199 ms | 11 - 367 MB | NPU
| detector | ONNX | mixed_with_float | Qualcomm® Dragonwing™ IQ-9075 | 15.669 ms | 6 - 14 MB | NPU
| detector | ONNX | mixed_with_float | Qualcomm® Dragonwing™ IQ-X7181 | 16.328 ms | 34 - 34 MB | NPU
| detector | ONNX | mixed_with_float | Qualcomm® Dragonwing™ Q-8750 | 9.35 ms | 7 - 281 MB | NPU
| detector | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 8.071 ms | 5 - 255 MB | NPU
| detector | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 11.174 ms | 0 - 248 MB | NPU
| detector | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 15.635 ms | 3 - 327 MB | NPU
| detector | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 76.316 ms | 4 - 321 MB | NPU
| detector | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 34.697 ms | 5 - 13 MB | NPU
| detector | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 19.67 ms | 5 - 7 MB | NPU
| detector | QNN_DLC | float | Qualcomm® SA8650P | 30.339 ms | 3 - 297 MB | NPU
| detector | QNN_DLC | float | Qualcomm® SA8255P | 30.339 ms | 3 - 297 MB | NPU
| detector | QNN_DLC | float | Qualcomm® QCS8450 | 76.316 ms | 4 - 321 MB | NPU
| detector | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 31.478 ms | 5 - 12 MB | NPU
| detector | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 11.174 ms | 0 - 248 MB | NPU
| detector | QNN_DLC | float | Qualcomm® SA8295P | 55.874 ms | 0 - 246 MB | NPU
| detector | QNN_DLC | mixed_with_float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 7.237 ms | 5 - 271 MB | NPU
| detector | QNN_DLC | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 9.124 ms | 0 - 260 MB | NPU
| detector | QNN_DLC | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 10.734 ms | 5 - 346 MB | NPU
| detector | QNN_DLC | mixed_with_float | Qualcomm® Dragonwing™ IQ-8275 | 15.191 ms | 5 - 13 MB | NPU
| detector | QNN_DLC | mixed_with_float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 15.285 ms | 5 - 11 MB | NPU
| detector | QNN_DLC | mixed_with_float | Qualcomm® SA8650P | 18.211 ms | 1 - 307 MB | NPU
| detector | QNN_DLC | mixed_with_float | Qualcomm® SA8255P | 18.211 ms | 1 - 307 MB | NPU
| detector | QNN_DLC | mixed_with_float | Qualcomm® Dragonwing™ IQ-9075 | 15.97 ms | 7 - 14 MB | NPU
| detector | QNN_DLC | mixed_with_float | Qualcomm® Dragonwing™ Q-8750 | 9.124 ms | 0 - 260 MB | NPU
| detector | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 9.305 ms | 0 - 256 MB | NPU
| detector | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 11.593 ms | 0 - 251 MB | NPU
| detector | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 16.661 ms | 0 - 335 MB | NPU
| detector | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 76.469 ms | 1 - 335 MB | NPU
| detector | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 36.514 ms | 0 - 84 MB | NPU
| detector | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 20.687 ms | 0 - 8 MB | NPU
| detector | TFLITE | float | Qualcomm® SA8650P | 31.601 ms | 0 - 267 MB | NPU
| detector | TFLITE | float | Qualcomm® SA8255P | 31.601 ms | 0 - 267 MB | NPU
| detector | TFLITE | float | Qualcomm® QCS8450 | 76.469 ms | 1 - 335 MB | NPU
| detector | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 30.763 ms | 0 - 83 MB | NPU
| detector | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 11.593 ms | 0 - 251 MB | NPU
| detector | TFLITE | float | Qualcomm® SA8295P | 57.059 ms | 0 - 254 MB | NPU
| text_encoder | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 31.029 ms | 1 - 2176 MB | NPU
| text_encoder | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 44.924 ms | 1 - 1981 MB | NPU
| text_encoder | ONNX | float | Snapdragon® X2 Elite | 33.125 ms | 1 - 1 MB | NPU
| text_encoder | ONNX | float | Snapdragon® X Elite | 83.669 ms | 551 - 551 MB | NPU
| text_encoder | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 60.151 ms | 1 - 3626 MB | NPU
| text_encoder | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 110.729 ms | 2 - 1814 MB | NPU
| text_encoder | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 92.427 ms | 0 - 4 MB | NPU
| text_encoder | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 80.593 ms | 0 - 644 MB | NPU
| text_encoder | ONNX | float | Qualcomm® QCS8450 | 110.729 ms | 2 - 1814 MB | NPU
| text_encoder | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 90.841 ms | 1 - 4 MB | NPU
| text_encoder | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 83.669 ms | 551 - 551 MB | NPU
| text_encoder | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 44.924 ms | 1 - 1981 MB | NPU
| text_encoder | ONNX | mixed_with_float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 31.158 ms | 1 - 2175 MB | NPU
| text_encoder | ONNX | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 44.333 ms | 1 - 1980 MB | NPU
| text_encoder | ONNX | mixed_with_float | Snapdragon® X2 Elite | 33.202 ms | 1 - 1 MB | NPU
| text_encoder | ONNX | mixed_with_float | Snapdragon® X Elite | 83.512 ms | 550 - 550 MB | NPU
| text_encoder | ONNX | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 60.158 ms | 1 - 3626 MB | NPU
| text_encoder | ONNX | mixed_with_float | Snapdragon® 8 Gen 1 Mobile | 111.209 ms | 2 - 1816 MB | NPU
| text_encoder | ONNX | mixed_with_float | Qualcomm® Dragonwing™ IQ-8275 | 92.349 ms | 0 - 4 MB | NPU
| text_encoder | ONNX | mixed_with_float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 80.847 ms | 0 - 644 MB | NPU
| text_encoder | ONNX | mixed_with_float | Qualcomm® QCS8450 | 111.209 ms | 2 - 1816 MB | NPU
| text_encoder | ONNX | mixed_with_float | Qualcomm® Dragonwing™ IQ-9075 | 90.752 ms | 1 - 4 MB | NPU
| text_encoder | ONNX | mixed_with_float | Qualcomm® Dragonwing™ IQ-X7181 | 83.512 ms | 550 - 550 MB | NPU
| text_encoder | ONNX | mixed_with_float | Qualcomm® Dragonwing™ Q-8750 | 44.333 ms | 1 - 1980 MB | NPU
| text_encoder | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 30.549 ms | 0 - 1923 MB | NPU
| text_encoder | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 43.419 ms | 0 - 2052 MB | NPU
| text_encoder | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 55.821 ms | 0 - 2293 MB | NPU
| text_encoder | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 106.955 ms | 0 - 1823 MB | NPU
| text_encoder | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 89.81 ms | 0 - 3 MB | NPU
| text_encoder | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 77.06 ms | 0 - 2 MB | NPU
| text_encoder | QNN_DLC | float | Qualcomm® SA8650P | 95.659 ms | 1 - 2019 MB | NPU
| text_encoder | QNN_DLC | float | Qualcomm® SA8255P | 95.659 ms | 1 - 2019 MB | NPU
| text_encoder | QNN_DLC | float | Qualcomm® QCS8450 | 106.955 ms | 0 - 1823 MB | NPU
| text_encoder | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 86.62 ms | 0 - 2 MB | NPU
| text_encoder | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 43.419 ms | 0 - 2052 MB | NPU
| text_encoder | QNN_DLC | float | Qualcomm® SA8295P | 102.822 ms | 0 - 1699 MB | NPU
| text_encoder | QNN_DLC | mixed_with_float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 30.66 ms | 0 - 1922 MB | NPU
| text_encoder | QNN_DLC | mixed_with_float | Snapdragon® 8 Elite For Galaxy Mobile | 43.864 ms | 0 - 2050 MB | NPU
| text_encoder | QNN_DLC | mixed_with_float | Snapdragon® 8 Gen 3 Mobile | 58.135 ms | 9 - 2303 MB | NPU
| text_encoder | QNN_DLC | mixed_with_float | Qualcomm® Dragonwing™ IQ-8275 | 90.055 ms | 0 - 3 MB | NPU
| text_encoder | QNN_DLC | mixed_with_float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 77.593 ms | 0 - 3 MB | NPU
| text_encoder | QNN_DLC | mixed_with_float | Qualcomm® SA8650P | 95.579 ms | 1 - 2019 MB | NPU
| text_encoder | QNN_DLC | mixed_with_float | Qualcomm® SA8255P | 95.579 ms | 1 - 2019 MB | NPU
| text_encoder | QNN_DLC | mixed_with_float | Qualcomm® Dragonwing™ IQ-9075 | 87.716 ms | 2 - 4 MB | NPU
| text_encoder | QNN_DLC | mixed_with_float | Qualcomm® Dragonwing™ Q-8750 | 43.864 ms | 0 - 2050 MB | NPU
| text_encoder | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 33.01 ms | 0 - 1764 MB | NPU
| text_encoder | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 50.301 ms | 0 - 1731 MB | NPU
| text_encoder | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 63.551 ms | 0 - 1978 MB | NPU
| text_encoder | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 112.967 ms | 1 - 1865 MB | NPU
| text_encoder | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 95.653 ms | 0 - 553 MB | NPU
| text_encoder | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 84.951 ms | 0 - 4 MB | NPU
| text_encoder | TFLITE | float | Qualcomm® SA8650P | 111.23 ms | 1 - 2145 MB | NPU
| text_encoder | TFLITE | float | Qualcomm® SA8255P | 111.23 ms | 1 - 2145 MB | NPU
| text_encoder | TFLITE | float | Qualcomm® QCS8450 | 112.967 ms | 1 - 1865 MB | NPU
| text_encoder | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 93.836 ms | 0 - 552 MB | NPU
| text_encoder | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 50.301 ms | 0 - 1731 MB | NPU
| text_encoder | TFLITE | float | Qualcomm® SA8295P | 111.099 ms | 1 - 1736 MB | NPU

## License
* The license for the original implementation of WeDetect can be found
  [here](https://github.com/WeChatCV/WeDetect/blob/main/README.md).

## References
* [WeDetect: Fast Open-Vocabulary Object Detection as Retrieval](https://arxiv.org/abs/2512.12309)
* [Source Model Implementation](https://github.com/WeChatCV/WeDetect)

## 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).