Hugging Face's logo

qualcomm
/
HRNetFace

Object Detection
PyTorch
real_time
android
Model card Files
xet
 Community
HRNetFace
File size: 10,040 Bytes
ba41173
 
 
 
 
 
 
 
 
 
 
 
1529439
ba41173
 
 
1529439
e4208df
1529439
 
 
 
 
 
 
 
 
 
 
 
3d435c2
 
 
 
 
 
1529439
 
 
 
 
 
e4208df
1529439
 
 
 
 
 
e4208df
1529439
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3d435c2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
caca897
3d435c2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ba41173
 
 
 
 
 
 
 
 
 
 
 
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
---
library_name: pytorch
license: other
tags:
- real_time
- android
pipeline_tag: object-detection

---

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

# HRNetFace: Optimized for Qualcomm Devices

Detects attributes (liveness, eye closeness, mask presence, glasses presence, sunglasses presence) that apply to a given face.

This is based on the implementation of HRNetFace found [here](https://github.com/HRNet/HRNet-Facial-Landmark-Detection).
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/main/src/qai_hub_models/models/hrnet_face) 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 1.24.3 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/hrnet_face/releases/v0.51.0/hrnet_face-onnx-float.zip)
| ONNX | w8a8 | Universal | QAIRT 2.42, ONNX Runtime 1.24.3 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/hrnet_face/releases/v0.51.0/hrnet_face-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/hrnet_face/releases/v0.51.0/hrnet_face-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/hrnet_face/releases/v0.51.0/hrnet_face-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/hrnet_face/releases/v0.51.0/hrnet_face-tflite-float.zip)
| TFLITE | w8a8 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/hrnet_face/releases/v0.51.0/hrnet_face-tflite-w8a8.zip)

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


### Option 2: Export with Custom Configurations

Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/hrnet_face) 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 [HRNetFace on GitHub](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/hrnet_face) for usage instructions.

## Model Details

**Model Type:** Model_use_case.object_detection

**Model Stats:**
- Model checkpoint: HR18-COFW.pth
- Input resolution: 256x256
- Number of parameters: 9.68M
- Model size (float): 36.87MB
- Model size (w8a8): 17.7 MB

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| HRNetFace | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.323 ms | 1 - 98 MB | NPU
| HRNetFace | ONNX | float | Snapdragon® X2 Elite | 1.527 ms | 31 - 31 MB | NPU
| HRNetFace | ONNX | float | Snapdragon® X Elite | 3.336 ms | 30 - 30 MB | NPU
| HRNetFace | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 2.153 ms | 0 - 155 MB | NPU
| HRNetFace | ONNX | float | Qualcomm® QCS8550 (Proxy) | 2.976 ms | 0 - 34 MB | NPU
| HRNetFace | ONNX | float | Qualcomm® QCS9075 | 4.752 ms | 2 - 4 MB | NPU
| HRNetFace | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.648 ms | 0 - 97 MB | NPU
| HRNetFace | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.588 ms | 0 - 105 MB | NPU
| HRNetFace | ONNX | w8a8 | Snapdragon® X2 Elite | 0.655 ms | 15 - 15 MB | NPU
| HRNetFace | ONNX | w8a8 | Snapdragon® X Elite | 1.543 ms | 15 - 15 MB | NPU
| HRNetFace | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.91 ms | 0 - 152 MB | NPU
| HRNetFace | ONNX | w8a8 | Qualcomm® QCS6490 | 91.333 ms | 18 - 36 MB | CPU
| HRNetFace | ONNX | w8a8 | Qualcomm® QCS8550 (Proxy) | 1.345 ms | 0 - 21 MB | NPU
| HRNetFace | ONNX | w8a8 | Qualcomm® QCS9075 | 1.53 ms | 0 - 3 MB | NPU
| HRNetFace | ONNX | w8a8 | Qualcomm® QCM6690 | 49.821 ms | 18 - 36 MB | CPU
| HRNetFace | ONNX | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.703 ms | 0 - 105 MB | NPU
| HRNetFace | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 46.758 ms | 17 - 35 MB | CPU
| HRNetFace | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.383 ms | 1 - 81 MB | NPU
| HRNetFace | QNN_DLC | float | Snapdragon® X2 Elite | 1.937 ms | 1 - 1 MB | NPU
| HRNetFace | QNN_DLC | float | Snapdragon® X Elite | 3.693 ms | 1 - 1 MB | NPU
| HRNetFace | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 2.305 ms | 1 - 122 MB | NPU
| HRNetFace | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 15.432 ms | 1 - 75 MB | NPU
| HRNetFace | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 3.276 ms | 1 - 3 MB | NPU
| HRNetFace | QNN_DLC | float | Qualcomm® SA8775P | 5.098 ms | 1 - 79 MB | NPU
| HRNetFace | QNN_DLC | float | Qualcomm® QCS9075 | 5.02 ms | 1 - 4 MB | NPU
| HRNetFace | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 5.502 ms | 0 - 104 MB | NPU
| HRNetFace | QNN_DLC | float | Qualcomm® SA7255P | 15.432 ms | 1 - 75 MB | NPU
| HRNetFace | QNN_DLC | float | Qualcomm® SA8295P | 5.64 ms | 0 - 62 MB | NPU
| HRNetFace | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.756 ms | 0 - 79 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.54 ms | 0 - 74 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 0.842 ms | 0 - 0 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Snapdragon® X Elite | 1.554 ms | 0 - 0 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.915 ms | 0 - 108 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Qualcomm® QCS6490 | 3.853 ms | 0 - 2 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Qualcomm® QCS8275 (Proxy) | 3.265 ms | 0 - 71 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Qualcomm® QCS8550 (Proxy) | 1.363 ms | 0 - 2 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Qualcomm® SA8775P | 1.716 ms | 0 - 72 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Qualcomm® QCS9075 | 1.611 ms | 0 - 2 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Qualcomm® QCM6690 | 10.627 ms | 0 - 200 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Qualcomm® QCS8450 (Proxy) | 1.841 ms | 0 - 108 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Qualcomm® SA7255P | 3.265 ms | 0 - 71 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Qualcomm® SA8295P | 2.222 ms | 0 - 70 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.69 ms | 0 - 74 MB | NPU
| HRNetFace | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 1.645 ms | 0 - 80 MB | NPU
| HRNetFace | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.378 ms | 0 - 96 MB | NPU
| HRNetFace | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 2.337 ms | 0 - 135 MB | NPU
| HRNetFace | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 15.489 ms | 0 - 93 MB | NPU
| HRNetFace | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 3.357 ms | 0 - 2 MB | NPU
| HRNetFace | TFLITE | float | Qualcomm® SA8775P | 5.129 ms | 0 - 92 MB | NPU
| HRNetFace | TFLITE | float | Qualcomm® QCS9075 | 5.052 ms | 0 - 35 MB | NPU
| HRNetFace | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 5.488 ms | 0 - 124 MB | NPU
| HRNetFace | TFLITE | float | Qualcomm® SA7255P | 15.489 ms | 0 - 93 MB | NPU
| HRNetFace | TFLITE | float | Qualcomm® SA8295P | 5.65 ms | 1 - 74 MB | NPU
| HRNetFace | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.765 ms | 0 - 96 MB | NPU
| HRNetFace | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.457 ms | 0 - 75 MB | NPU
| HRNetFace | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.752 ms | 0 - 110 MB | NPU
| HRNetFace | TFLITE | w8a8 | Qualcomm® QCS6490 | 3.357 ms | 0 - 18 MB | NPU
| HRNetFace | TFLITE | w8a8 | Qualcomm® QCS8275 (Proxy) | 2.912 ms | 0 - 72 MB | NPU
| HRNetFace | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 1.134 ms | 0 - 3 MB | NPU
| HRNetFace | TFLITE | w8a8 | Qualcomm® SA8775P | 1.482 ms | 0 - 74 MB | NPU
| HRNetFace | TFLITE | w8a8 | Qualcomm® QCS9075 | 1.303 ms | 0 - 18 MB | NPU
| HRNetFace | TFLITE | w8a8 | Qualcomm® QCM6690 | 9.846 ms | 0 - 193 MB | NPU
| HRNetFace | TFLITE | w8a8 | Qualcomm® QCS8450 (Proxy) | 1.557 ms | 0 - 112 MB | NPU
| HRNetFace | TFLITE | w8a8 | Qualcomm® SA7255P | 2.912 ms | 0 - 72 MB | NPU
| HRNetFace | TFLITE | w8a8 | Qualcomm® SA8295P | 1.989 ms | 0 - 70 MB | NPU
| HRNetFace | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.578 ms | 0 - 71 MB | NPU
| HRNetFace | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 1.352 ms | 0 - 75 MB | NPU

## License
* The license for the original implementation of HRNetFace can be found
  [here](https://github.com/HRNet/HRNet-Facial-Landmark-Detection/blob/master/LICENCE).

## References
* [Deep High-Resolution Representation Learning for Visual Recognition](https://arxiv.org/abs/1908.07919)
* [Source Model Implementation](https://github.com/HRNet/HRNet-Facial-Landmark-Detection)

## 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).