Hugging Face's logo

qualcomm
/
SixDRepNet

Keypoint Detection
PyTorch
real_time
android
Model card Files
xet
 Community
SixDRepNet
File size: 10,557 Bytes
60231d7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c12476a
 
 
60231d7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
library_name: pytorch
license: other
tags:
- real_time
- android
pipeline_tag: keypoint-detection

---

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

# SixDRepNet: Optimized for Qualcomm Devices

6DRepNet predicts head pose (pitch, yaw, roll) from a face image using a RepVGG-B1g2 backbone and a continuous 6D rotation representation, achieving robust and accurate head pose estimation.

This is based on the implementation of SixDRepNet found [here](https://github.com/thohemp/6DRepNet).
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/sixd_repnet) 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/sixd_repnet/releases/v0.53.1/sixd_repnet-onnx-float.zip)
| QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sixd_repnet/releases/v0.53.1/sixd_repnet-qnn_dlc-float.zip)
| TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sixd_repnet/releases/v0.53.1/sixd_repnet-tflite-float.zip)

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


### 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/sixd_repnet) 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 [SixDRepNet on GitHub](https://github.com/qualcomm/ai-hub-models/blob/main/src/qai_hub_models/models/sixd_repnet) for usage instructions.

## Model Details

**Model Type:** Model_use_case.pose_estimation

**Model Stats:**
- Input resolution: 224x224
- Number of parameters: 15.3M
- Model size (float): 58.4 MB

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| face_detector | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.483 ms | 0 - 166 MB | NPU
| face_detector | ONNX | float | Snapdragon® 8 Elite Mobile | 1.863 ms | 3 - 166 MB | NPU
| face_detector | ONNX | float | Snapdragon® X2 Elite | 1.581 ms | 7 - 7 MB | NPU
| face_detector | ONNX | float | Snapdragon® X Elite | 3.803 ms | 7 - 7 MB | NPU
| face_detector | ONNX | float | Snapdragon® X Elite | 3.803 ms | 7 - 7 MB | NPU
| face_detector | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 2.201 ms | 3 - 169 MB | NPU
| face_detector | ONNX | float | Qualcomm® QCS8550 (Proxy) | 3.469 ms | 0 - 4 MB | NPU
| face_detector | ONNX | float | Qualcomm® QCS9075 | 5.366 ms | 4 - 12 MB | NPU
| face_detector | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.863 ms | 3 - 166 MB | NPU
| face_detector | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 5.42 ms | 5 - 159 MB | NPU
| face_detector | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 6.896 ms | 0 - 150 MB | NPU
| face_detector | QNN_DLC | float | Snapdragon® X2 Elite | 5.876 ms | 5 - 5 MB | NPU
| face_detector | QNN_DLC | float | Snapdragon® X Elite | 16.475 ms | 5 - 5 MB | NPU
| face_detector | QNN_DLC | float | Snapdragon® X Elite | 16.475 ms | 5 - 5 MB | NPU
| face_detector | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 9.278 ms | 5 - 173 MB | NPU
| face_detector | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 28.312 ms | 1 - 151 MB | NPU
| face_detector | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 15.67 ms | 5 - 8 MB | NPU
| face_detector | QNN_DLC | float | Qualcomm® SA8775P | 16.54 ms | 1 - 153 MB | NPU
| face_detector | QNN_DLC | float | Qualcomm® SA8775P | 16.54 ms | 1 - 153 MB | NPU
| face_detector | QNN_DLC | float | Qualcomm® SA8775P | 16.54 ms | 1 - 153 MB | NPU
| face_detector | QNN_DLC | float | Qualcomm® QCS9075 | 19.498 ms | 5 - 12 MB | NPU
| face_detector | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 24.733 ms | 5 - 179 MB | NPU
| face_detector | QNN_DLC | float | Qualcomm® SA7255P | 28.312 ms | 1 - 151 MB | NPU
| face_detector | QNN_DLC | float | Qualcomm® SA8295P | 20.592 ms | 0 - 150 MB | NPU
| face_detector | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 6.896 ms | 0 - 150 MB | NPU
| face_detector | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 5.482 ms | 1 - 155 MB | NPU
| face_detector | TFLITE | float | Snapdragon® 8 Elite Mobile | 6.923 ms | 0 - 150 MB | NPU
| face_detector | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 9.272 ms | 1 - 171 MB | NPU
| face_detector | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 28.26 ms | 1 - 150 MB | NPU
| face_detector | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 15.666 ms | 1 - 3 MB | NPU
| face_detector | TFLITE | float | Qualcomm® SA8775P | 16.531 ms | 1 - 152 MB | NPU
| face_detector | TFLITE | float | Qualcomm® SA8775P | 16.531 ms | 1 - 152 MB | NPU
| face_detector | TFLITE | float | Qualcomm® SA8775P | 16.531 ms | 1 - 152 MB | NPU
| face_detector | TFLITE | float | Qualcomm® QCS9075 | 19.66 ms | 1 - 10 MB | NPU
| face_detector | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 24.453 ms | 1 - 171 MB | NPU
| face_detector | TFLITE | float | Qualcomm® SA7255P | 28.26 ms | 1 - 150 MB | NPU
| face_detector | TFLITE | float | Qualcomm® SA8295P | 20.653 ms | 1 - 152 MB | NPU
| face_detector | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 6.923 ms | 0 - 150 MB | NPU
| pose_estimator | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.334 ms | 0 - 24 MB | NPU
| pose_estimator | ONNX | float | Snapdragon® 8 Elite Mobile | 1.619 ms | 0 - 22 MB | NPU
| pose_estimator | ONNX | float | Snapdragon® X2 Elite | 1.318 ms | 75 - 75 MB | NPU
| pose_estimator | ONNX | float | Snapdragon® X Elite | 2.633 ms | 75 - 75 MB | NPU
| pose_estimator | ONNX | float | Snapdragon® X Elite | 2.633 ms | 75 - 75 MB | NPU
| pose_estimator | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 2.059 ms | 0 - 35 MB | NPU
| pose_estimator | ONNX | float | Qualcomm® QCS8550 (Proxy) | 2.627 ms | 1 - 2 MB | NPU
| pose_estimator | ONNX | float | Qualcomm® QCS9075 | 4.551 ms | 0 - 4 MB | NPU
| pose_estimator | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.619 ms | 0 - 22 MB | NPU
| pose_estimator | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.417 ms | 1 - 29 MB | NPU
| pose_estimator | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 1.682 ms | 0 - 26 MB | NPU
| pose_estimator | QNN_DLC | float | Snapdragon® X2 Elite | 1.487 ms | 1 - 1 MB | NPU
| pose_estimator | QNN_DLC | float | Snapdragon® X Elite | 2.864 ms | 1 - 1 MB | NPU
| pose_estimator | QNN_DLC | float | Snapdragon® X Elite | 2.864 ms | 1 - 1 MB | NPU
| pose_estimator | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 2.235 ms | 1 - 40 MB | NPU
| pose_estimator | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 17.816 ms | 1 - 24 MB | NPU
| pose_estimator | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 2.782 ms | 1 - 2 MB | NPU
| pose_estimator | QNN_DLC | float | Qualcomm® SA8775P | 4.813 ms | 1 - 26 MB | NPU
| pose_estimator | QNN_DLC | float | Qualcomm® SA8775P | 4.813 ms | 1 - 26 MB | NPU
| pose_estimator | QNN_DLC | float | Qualcomm® SA8775P | 4.813 ms | 1 - 26 MB | NPU
| pose_estimator | QNN_DLC | float | Qualcomm® QCS9075 | 4.904 ms | 3 - 5 MB | NPU
| pose_estimator | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 6.543 ms | 0 - 39 MB | NPU
| pose_estimator | QNN_DLC | float | Qualcomm® SA7255P | 17.816 ms | 1 - 24 MB | NPU
| pose_estimator | QNN_DLC | float | Qualcomm® SA8295P | 5.398 ms | 1 - 23 MB | NPU
| pose_estimator | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.682 ms | 0 - 26 MB | NPU
| pose_estimator | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.374 ms | 0 - 28 MB | NPU
| pose_estimator | TFLITE | float | Snapdragon® 8 Elite Mobile | 1.701 ms | 0 - 30 MB | NPU
| pose_estimator | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 2.265 ms | 0 - 46 MB | NPU
| pose_estimator | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 17.398 ms | 0 - 26 MB | NPU
| pose_estimator | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 2.849 ms | 0 - 3 MB | NPU
| pose_estimator | TFLITE | float | Qualcomm® SA8775P | 4.813 ms | 0 - 27 MB | NPU
| pose_estimator | TFLITE | float | Qualcomm® SA8775P | 4.813 ms | 0 - 27 MB | NPU
| pose_estimator | TFLITE | float | Qualcomm® SA8775P | 4.813 ms | 0 - 27 MB | NPU
| pose_estimator | TFLITE | float | Qualcomm® QCS9075 | 4.74 ms | 0 - 78 MB | NPU
| pose_estimator | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 6.502 ms | 0 - 43 MB | NPU
| pose_estimator | TFLITE | float | Qualcomm® SA7255P | 17.398 ms | 0 - 26 MB | NPU
| pose_estimator | TFLITE | float | Qualcomm® SA8295P | 5.349 ms | 0 - 28 MB | NPU
| pose_estimator | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.701 ms | 0 - 30 MB | NPU

## License
* The license for the original implementation of SixDRepNet can be found
  [here](https://github.com/thohemp/6DRepNet/blob/master/LICENSE).

## References
* [6D Rotation Representation for Unconstrained Head Pose Estimation](https://arxiv.org/abs/2109.10948)
* [Source Model Implementation](https://github.com/thohemp/6DRepNet)

## 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).