Hugging Face's logo

qualcomm
/
Facial-Landmark-Detection

Keypoint Detection
PyTorch
backbone
android
Model card Files
xet
Community
Facial-Landmark-Detection
File size: 10,871 Bytes
e51ed16
 
80261e3
e51ed16
 
 
ab12950
e51ed16
 
 
 
 
e73bc40
6460062
63571a0
6460062
e73bc40
 
 
 
 
 
 
 
 
 
 
 
 
3392edc
 
 
 
 
 
e73bc40
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3392edc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e73bc40
e51ed16
 
2484f1b
b1160d9
e51ed16
 
 
 
 
 
 
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:
- backbone
- android
pipeline_tag: keypoint-detection

---

![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/facemap_3dmm/web-assets/model_demo.png)

# Facial-Landmark-Detection: Optimized for Qualcomm Devices

Detects facial landmarks (eg, nose, mouth, etc.). This model's architecture was developed by Qualcomm. The model was trained by Qualcomm on a proprietary dataset of faces, but can be used on any image.

This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/facemap_3dmm) 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/facemap_3dmm/releases/v0.47.0/facemap_3dmm-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/facemap_3dmm/releases/v0.47.0/facemap_3dmm-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/facemap_3dmm/releases/v0.47.0/facemap_3dmm-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/facemap_3dmm/releases/v0.47.0/facemap_3dmm-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/facemap_3dmm/releases/v0.47.0/facemap_3dmm-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/facemap_3dmm/releases/v0.47.0/facemap_3dmm-tflite-w8a8.zip)

For more device-specific assets and performance metrics, visit **[Facial-Landmark-Detection on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/facemap_3dmm)**.


### Option 2: Export with Custom Configurations

Use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/facemap_3dmm) 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 [Facial-Landmark-Detection on GitHub](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/facemap_3dmm) for usage instructions.

## Model Details

**Model Type:** Model_use_case.pose_estimation

**Model Stats:**
- Input resolution: 128x128
- Number of parameters: 5.42M
- Model size (float): 20.7 MB
- Model size (w8a8): 5.27 MB

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| Facial-Landmark-Detection | ONNX | float | Snapdragon® X Elite | 0.394 ms | 10 - 10 MB | NPU
| Facial-Landmark-Detection | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.241 ms | 0 - 29 MB | NPU
| Facial-Landmark-Detection | ONNX | float | Qualcomm® QCS8550 (Proxy) | 0.32 ms | 0 - 32 MB | NPU
| Facial-Landmark-Detection | ONNX | float | Qualcomm® QCS9075 | 0.436 ms | 0 - 3 MB | NPU
| Facial-Landmark-Detection | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.225 ms | 0 - 24 MB | NPU
| Facial-Landmark-Detection | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.208 ms | 0 - 23 MB | NPU
| Facial-Landmark-Detection | ONNX | float | Snapdragon® X2 Elite | 0.168 ms | 10 - 10 MB | NPU
| Facial-Landmark-Detection | ONNX | w8a8 | Snapdragon® X Elite | 0.237 ms | 5 - 5 MB | NPU
| Facial-Landmark-Detection | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.139 ms | 0 - 38 MB | NPU
| Facial-Landmark-Detection | ONNX | w8a8 | Qualcomm® QCS6490 | 3.099 ms | 1 - 8 MB | CPU
| Facial-Landmark-Detection | ONNX | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.173 ms | 0 - 8 MB | NPU
| Facial-Landmark-Detection | ONNX | w8a8 | Qualcomm® QCS9075 | 0.247 ms | 0 - 3 MB | NPU
| Facial-Landmark-Detection | ONNX | w8a8 | Qualcomm® QCM6690 | 1.684 ms | 0 - 6 MB | CPU
| Facial-Landmark-Detection | ONNX | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.12 ms | 0 - 20 MB | NPU
| Facial-Landmark-Detection | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 1.271 ms | 0 - 7 MB | CPU
| Facial-Landmark-Detection | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.111 ms | 0 - 23 MB | NPU
| Facial-Landmark-Detection | ONNX | w8a8 | Snapdragon® X2 Elite | 0.077 ms | 5 - 5 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Snapdragon® X Elite | 0.374 ms | 0 - 0 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 0.223 ms | 0 - 30 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 1.123 ms | 0 - 21 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 0.294 ms | 0 - 2 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Qualcomm® SA8775P | 0.5 ms | 0 - 22 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Qualcomm® QCS9075 | 0.391 ms | 0 - 2 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 0.549 ms | 0 - 30 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Qualcomm® SA7255P | 1.123 ms | 0 - 21 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Qualcomm® SA8295P | 0.688 ms | 0 - 17 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.197 ms | 0 - 20 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.189 ms | 0 - 23 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | float | Snapdragon® X2 Elite | 0.246 ms | 0 - 0 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Snapdragon® X Elite | 0.237 ms | 0 - 0 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.14 ms | 0 - 36 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Qualcomm® QCS6490 | 0.708 ms | 0 - 2 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Qualcomm® QCS8275 (Proxy) | 0.453 ms | 0 - 20 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.163 ms | 0 - 1 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Qualcomm® SA8775P | 0.323 ms | 0 - 21 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Qualcomm® QCS9075 | 0.225 ms | 0 - 2 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Qualcomm® QCM6690 | 0.595 ms | 0 - 23 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Qualcomm® QCS8450 (Proxy) | 0.238 ms | 0 - 37 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Qualcomm® SA7255P | 0.453 ms | 0 - 20 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Qualcomm® SA8295P | 0.457 ms | 0 - 17 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.113 ms | 0 - 19 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.214 ms | 2 - 25 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.11 ms | 0 - 22 MB | NPU
| Facial-Landmark-Detection | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 0.166 ms | 0 - 0 MB | NPU
| Facial-Landmark-Detection | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 0.224 ms | 0 - 41 MB | NPU
| Facial-Landmark-Detection | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 1.156 ms | 0 - 22 MB | NPU
| Facial-Landmark-Detection | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 0.282 ms | 0 - 2 MB | NPU
| Facial-Landmark-Detection | TFLITE | float | Qualcomm® SA8775P | 0.498 ms | 0 - 24 MB | NPU
| Facial-Landmark-Detection | TFLITE | float | Qualcomm® QCS9075 | 0.382 ms | 0 - 12 MB | NPU
| Facial-Landmark-Detection | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 0.474 ms | 0 - 41 MB | NPU
| Facial-Landmark-Detection | TFLITE | float | Qualcomm® SA7255P | 1.156 ms | 0 - 22 MB | NPU
| Facial-Landmark-Detection | TFLITE | float | Qualcomm® SA8295P | 0.682 ms | 0 - 18 MB | NPU
| Facial-Landmark-Detection | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.195 ms | 0 - 21 MB | NPU
| Facial-Landmark-Detection | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.192 ms | 0 - 24 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.136 ms | 0 - 36 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Qualcomm® QCS6490 | 0.59 ms | 0 - 7 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Qualcomm® QCS8275 (Proxy) | 0.483 ms | 0 - 20 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.169 ms | 0 - 1 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Qualcomm® SA8775P | 0.335 ms | 0 - 21 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Qualcomm® QCS9075 | 0.227 ms | 0 - 7 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Qualcomm® QCM6690 | 0.599 ms | 0 - 22 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Qualcomm® QCS8450 (Proxy) | 0.239 ms | 0 - 37 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Qualcomm® SA7255P | 0.483 ms | 0 - 20 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Qualcomm® SA8295P | 0.485 ms | 0 - 17 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.123 ms | 0 - 18 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.219 ms | 0 - 23 MB | NPU
| Facial-Landmark-Detection | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.118 ms | 0 - 22 MB | NPU

## License
* The license for the original implementation of Facial-Landmark-Detection can be found
  [here](https://github.com/quic/ai-hub-models/blob/main/LICENSE).



## 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).