Hugging Face's logo

qualcomm
/
SixDRepNet

Keypoint Detection
PyTorch
real_time
android
Model card Files
xet
 Community
qaihm-bot commited on
Commit
60231d7
·
verified ·
1 Parent(s): b290952

v0.53.0

Browse files

See https://github.com/qualcomm/ai-hub-models/releases/v0.53.0 for changelog.

Files changed (3) hide show
  1. LICENSE +1 -0
  2. README.md +148 -0
  3. release_assets.json +29 -0
LICENSE ADDED
@@ -0,0 +1 @@
 
 
1
+ The license of the original trained model can be found at https://github.com/thohemp/6DRepNet/blob/master/LICENSE.
README.md ADDED
@@ -0,0 +1,148 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: pytorch
3
+ license: other
4
+ tags:
5
+ - real_time
6
+ - android
7
+ pipeline_tag: keypoint-detection
8
+
9
+ ---
10
+
11
+ ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sixd_repnet/web-assets/model_demo.png)
12
+
13
+ # SixDRepNet: Optimized for Qualcomm Devices
14
+
15
+ 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.
16
+
17
+ This is based on the implementation of SixDRepNet found [here](https://github.com/thohemp/6DRepNet).
18
+ 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).
19
+
20
+ 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.
21
+
22
+ ## Getting Started
23
+ There are two ways to deploy this model on your device:
24
+
25
+ ### Option 1: Download Pre-Exported Models
26
+
27
+ Below are pre-exported model assets ready for deployment.
28
+
29
+ | Runtime | Precision | Chipset | SDK Versions | Download |
30
+ |---|---|---|---|---|
31
+ | 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/v0.53.0/sixd_repnet-onnx-float.zip)
32
+ | 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/v0.53.0/sixd_repnet-qnn_dlc-float.zip)
33
+ | TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sixd_repnet/releases/v0.53.0/sixd_repnet-tflite-float.zip)
34
+
35
+ For more device-specific assets and performance metrics, visit **[SixDRepNet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/sixd_repnet)**.
36
+
37
+
38
+ ### Option 2: Export with Custom Configurations
39
+
40
+ 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:
41
+ - Custom weights (e.g., fine-tuned checkpoints)
42
+ - Custom input shapes
43
+ - Target device and runtime configurations
44
+
45
+ This option is ideal if you need to customize the model beyond the default configuration provided here.
46
+
47
+ 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.
48
+
49
+ ## Model Details
50
+
51
+ **Model Type:** Model_use_case.pose_estimation
52
+
53
+ **Model Stats:**
54
+ - Input resolution: 224x224
55
+ - Number of parameters: 15.3M
56
+ - Model size (float): 58.4 MB
57
+
58
+ ## Performance Summary
59
+ | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
60
+ |---|---|---|---|---|---|---
61
+ | face_detector | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.483 ms | 0 - 166 MB | NPU
62
+ | face_detector | ONNX | float | Snapdragon® 8 Elite Mobile | 1.863 ms | 3 - 166 MB | NPU
63
+ | face_detector | ONNX | float | Snapdragon® X2 Elite | 1.581 ms | 7 - 7 MB | NPU
64
+ | face_detector | ONNX | float | Snapdragon® X Elite | 3.803 ms | 7 - 7 MB | NPU
65
+ | face_detector | ONNX | float | Snapdragon® X Elite | 3.803 ms | 7 - 7 MB | NPU
66
+ | face_detector | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 2.201 ms | 3 - 169 MB | NPU
67
+ | face_detector | ONNX | float | Qualcomm® QCS8550 (Proxy) | 3.469 ms | 0 - 4 MB | NPU
68
+ | face_detector | ONNX | float | Qualcomm® QCS9075 | 5.366 ms | 4 - 12 MB | NPU
69
+ | face_detector | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.863 ms | 3 - 166 MB | NPU
70
+ | face_detector | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 5.42 ms | 5 - 159 MB | NPU
71
+ | face_detector | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 6.896 ms | 0 - 150 MB | NPU
72
+ | face_detector | QNN_DLC | float | Snapdragon® X2 Elite | 5.876 ms | 5 - 5 MB | NPU
73
+ | face_detector | QNN_DLC | float | Snapdragon® X Elite | 16.475 ms | 5 - 5 MB | NPU
74
+ | face_detector | QNN_DLC | float | Snapdragon® X Elite | 16.475 ms | 5 - 5 MB | NPU
75
+ | face_detector | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 9.278 ms | 5 - 173 MB | NPU
76
+ | face_detector | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 28.312 ms | 1 - 151 MB | NPU
77
+ | face_detector | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 15.67 ms | 5 - 8 MB | NPU
78
+ | face_detector | QNN_DLC | float | Qualcomm® SA8775P | 16.54 ms | 1 - 153 MB | NPU
79
+ | face_detector | QNN_DLC | float | Qualcomm® SA8775P | 16.54 ms | 1 - 153 MB | NPU
80
+ | face_detector | QNN_DLC | float | Qualcomm® SA8775P | 16.54 ms | 1 - 153 MB | NPU
81
+ | face_detector | QNN_DLC | float | Qualcomm® QCS9075 | 19.498 ms | 5 - 12 MB | NPU
82
+ | face_detector | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 24.733 ms | 5 - 179 MB | NPU
83
+ | face_detector | QNN_DLC | float | Qualcomm® SA7255P | 28.312 ms | 1 - 151 MB | NPU
84
+ | face_detector | QNN_DLC | float | Qualcomm® SA8295P | 20.592 ms | 0 - 150 MB | NPU
85
+ | face_detector | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 6.896 ms | 0 - 150 MB | NPU
86
+ | face_detector | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 5.482 ms | 1 - 155 MB | NPU
87
+ | face_detector | TFLITE | float | Snapdragon® 8 Elite Mobile | 6.923 ms | 0 - 150 MB | NPU
88
+ | face_detector | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 9.272 ms | 1 - 171 MB | NPU
89
+ | face_detector | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 28.26 ms | 1 - 150 MB | NPU
90
+ | face_detector | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 15.666 ms | 1 - 3 MB | NPU
91
+ | face_detector | TFLITE | float | Qualcomm® SA8775P | 16.531 ms | 1 - 152 MB | NPU
92
+ | face_detector | TFLITE | float | Qualcomm® SA8775P | 16.531 ms | 1 - 152 MB | NPU
93
+ | face_detector | TFLITE | float | Qualcomm® SA8775P | 16.531 ms | 1 - 152 MB | NPU
94
+ | face_detector | TFLITE | float | Qualcomm® QCS9075 | 19.66 ms | 1 - 10 MB | NPU
95
+ | face_detector | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 24.453 ms | 1 - 171 MB | NPU
96
+ | face_detector | TFLITE | float | Qualcomm® SA7255P | 28.26 ms | 1 - 150 MB | NPU
97
+ | face_detector | TFLITE | float | Qualcomm® SA8295P | 20.653 ms | 1 - 152 MB | NPU
98
+ | face_detector | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 6.923 ms | 0 - 150 MB | NPU
99
+ | pose_estimator | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.334 ms | 0 - 24 MB | NPU
100
+ | pose_estimator | ONNX | float | Snapdragon® 8 Elite Mobile | 1.619 ms | 0 - 22 MB | NPU
101
+ | pose_estimator | ONNX | float | Snapdragon® X2 Elite | 1.318 ms | 75 - 75 MB | NPU
102
+ | pose_estimator | ONNX | float | Snapdragon® X Elite | 2.633 ms | 75 - 75 MB | NPU
103
+ | pose_estimator | ONNX | float | Snapdragon® X Elite | 2.633 ms | 75 - 75 MB | NPU
104
+ | pose_estimator | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 2.059 ms | 0 - 35 MB | NPU
105
+ | pose_estimator | ONNX | float | Qualcomm® QCS8550 (Proxy) | 2.627 ms | 1 - 2 MB | NPU
106
+ | pose_estimator | ONNX | float | Qualcomm® QCS9075 | 4.551 ms | 0 - 4 MB | NPU
107
+ | pose_estimator | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.619 ms | 0 - 22 MB | NPU
108
+ | pose_estimator | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.417 ms | 1 - 29 MB | NPU
109
+ | pose_estimator | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 1.682 ms | 0 - 26 MB | NPU
110
+ | pose_estimator | QNN_DLC | float | Snapdragon® X2 Elite | 1.487 ms | 1 - 1 MB | NPU
111
+ | pose_estimator | QNN_DLC | float | Snapdragon® X Elite | 2.864 ms | 1 - 1 MB | NPU
112
+ | pose_estimator | QNN_DLC | float | Snapdragon® X Elite | 2.864 ms | 1 - 1 MB | NPU
113
+ | pose_estimator | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 2.235 ms | 1 - 40 MB | NPU
114
+ | pose_estimator | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 17.816 ms | 1 - 24 MB | NPU
115
+ | pose_estimator | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 2.782 ms | 1 - 2 MB | NPU
116
+ | pose_estimator | QNN_DLC | float | Qualcomm® SA8775P | 4.813 ms | 1 - 26 MB | NPU
117
+ | pose_estimator | QNN_DLC | float | Qualcomm® SA8775P | 4.813 ms | 1 - 26 MB | NPU
118
+ | pose_estimator | QNN_DLC | float | Qualcomm® SA8775P | 4.813 ms | 1 - 26 MB | NPU
119
+ | pose_estimator | QNN_DLC | float | Qualcomm® QCS9075 | 4.904 ms | 3 - 5 MB | NPU
120
+ | pose_estimator | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 6.543 ms | 0 - 39 MB | NPU
121
+ | pose_estimator | QNN_DLC | float | Qualcomm® SA7255P | 17.816 ms | 1 - 24 MB | NPU
122
+ | pose_estimator | QNN_DLC | float | Qualcomm® SA8295P | 5.398 ms | 1 - 23 MB | NPU
123
+ | pose_estimator | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.682 ms | 0 - 26 MB | NPU
124
+ | pose_estimator | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 1.374 ms | 0 - 28 MB | NPU
125
+ | pose_estimator | TFLITE | float | Snapdragon® 8 Elite Mobile | 1.701 ms | 0 - 30 MB | NPU
126
+ | pose_estimator | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 2.265 ms | 0 - 46 MB | NPU
127
+ | pose_estimator | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 17.398 ms | 0 - 26 MB | NPU
128
+ | pose_estimator | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 2.849 ms | 0 - 3 MB | NPU
129
+ | pose_estimator | TFLITE | float | Qualcomm® SA8775P | 4.813 ms | 0 - 27 MB | NPU
130
+ | pose_estimator | TFLITE | float | Qualcomm® SA8775P | 4.813 ms | 0 - 27 MB | NPU
131
+ | pose_estimator | TFLITE | float | Qualcomm® SA8775P | 4.813 ms | 0 - 27 MB | NPU
132
+ | pose_estimator | TFLITE | float | Qualcomm® QCS9075 | 4.74 ms | 0 - 78 MB | NPU
133
+ | pose_estimator | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 6.502 ms | 0 - 43 MB | NPU
134
+ | pose_estimator | TFLITE | float | Qualcomm® SA7255P | 17.398 ms | 0 - 26 MB | NPU
135
+ | pose_estimator | TFLITE | float | Qualcomm® SA8295P | 5.349 ms | 0 - 28 MB | NPU
136
+ | pose_estimator | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.701 ms | 0 - 30 MB | NPU
137
+
138
+ ## License
139
+ * The license for the original implementation of SixDRepNet can be found
140
+ [here](https://github.com/thohemp/6DRepNet/blob/master/LICENSE).
141
+
142
+ ## References
143
+ * [6D Rotation Representation for Unconstrained Head Pose Estimation](https://arxiv.org/abs/2109.10948)
144
+ * [Source Model Implementation](https://github.com/thohemp/6DRepNet)
145
+
146
+ ## Community
147
+ * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
148
+ * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).
release_assets.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "0.53.0",
3
+ "precisions": {
4
+ "float": {
5
+ "universal_assets": {
6
+ "tflite": {
7
+ "tool_versions": {
8
+ "qairt": "2.45.0.260326154327",
9
+ "litert": "1.4.3"
10
+ },
11
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sixd_repnet/releases/v0.53.0/sixd_repnet-tflite-float.zip"
12
+ },
13
+ "qnn_dlc": {
14
+ "tool_versions": {
15
+ "qairt": "2.45.0.260326154327"
16
+ },
17
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sixd_repnet/releases/v0.53.0/sixd_repnet-qnn_dlc-float.zip"
18
+ },
19
+ "onnx": {
20
+ "tool_versions": {
21
+ "qairt": "2.42.0.251225135753_193295",
22
+ "onnx_runtime": "1.24.3"
23
+ },
24
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sixd_repnet/releases/v0.53.0/sixd_repnet-onnx-float.zip"
25
+ }
26
+ }
27
+ }
28
+ }
29
+ }