Hugging Face's logo

qualcomm
/
MobileFaceNet

Object Detection
PyTorch
real_time
bu_auto
android
Model card Files
xet
 Community
qaihm-bot commited on
Commit
c9c0357
·
verified ·
1 Parent(s): 80b4d22

v0.61.0

Browse files

See https://github.com/qualcomm/ai-hub-models/releases/v0.61.0 for changelog.

Files changed (3) hide show
  1. LICENSE +1 -0
  2. README.md +159 -0
  3. release_assets.json +45 -0
LICENSE ADDED
@@ -0,0 +1 @@
 
 
1
+ The license of the original trained model can be found at https://github.com/foamliu/MobileFaceNet/blob/master/LICENSE.
README.md ADDED
@@ -0,0 +1,159 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: pytorch
3
+ license: apache-2.0
4
+ tags:
5
+ - real_time
6
+ - bu_auto
7
+ - android
8
+ pipeline_tag: object-detection
9
+
10
+ ---
11
+
12
+ ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/mobile_facenet/web-assets/model_demo.png)
13
+
14
+ # MobileFaceNet: Optimized for Qualcomm Devices
15
+
16
+ MobileFaceNet is an efficient CNN that maps a 112x112 face image to a compact 128-dimensional embedding. Two embeddings are compared via cosine similarity to determine whether they belong to the same person, achieving 99.48% accuracy on the LFW benchmark. The model uses depthwise-separable convolutions and inverted residual blocks (MobileNetV2-style) to stay under 1M parameters, making it well-suited for real-time face verification on mobile and edge devices. Trained with ArcFace loss on MS-Celeb-1M.
17
+
18
+ This is based on the implementation of MobileFaceNet found [here](https://github.com/foamliu/MobileFaceNet).
19
+ 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/v0.61.0/src/qai_hub_models/models/mobile_facenet) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
20
+
21
+ 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.
22
+
23
+ ## Getting Started
24
+ There are two ways to deploy this model on your device:
25
+
26
+ ### Option 1: Download Pre-Exported Models
27
+
28
+ Below are pre-exported model assets ready for deployment.
29
+
30
+ | Runtime | Precision | Chipset | SDK Versions | Download |
31
+ |---|---|---|---|---|
32
+ | ONNX | float | Universal | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/mobile_facenet/releases/v0.61.0/mobile_facenet-onnx-float.zip)
33
+ | ONNX | w8a16 | Universal | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/mobile_facenet/releases/v0.61.0/mobile_facenet-onnx-w8a16.zip)
34
+ | QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/mobile_facenet/releases/v0.61.0/mobile_facenet-qnn_dlc-float.zip)
35
+ | QNN_DLC | w8a16 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/mobile_facenet/releases/v0.61.0/mobile_facenet-qnn_dlc-w8a16.zip)
36
+ | TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/mobile_facenet/releases/v0.61.0/mobile_facenet-tflite-float.zip)
37
+
38
+ For more device-specific assets and performance metrics, visit **[MobileFaceNet on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/mobile_facenet)**.
39
+
40
+
41
+ ### Option 2: Export with Custom Configurations
42
+
43
+ Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.61.0/src/qai_hub_models/models/mobile_facenet) Python library to compile and export the model with your own:
44
+ - Custom weights (e.g., fine-tuned checkpoints)
45
+ - Custom input shapes
46
+ - Target device and runtime configurations
47
+
48
+ This option is ideal if you need to customize the model beyond the default configuration provided here.
49
+
50
+ See our repository for [MobileFaceNet on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.61.0/src/qai_hub_models/models/mobile_facenet) for usage instructions.
51
+
52
+ ## Model Details
53
+
54
+ **Model Type:** Model_use_case.object_detection
55
+
56
+ **Model Stats:**
57
+ - Embedding dimension: 128
58
+ - Input resolution: 112x112
59
+ - Model checkpoint: mobilefacenet.pt
60
+ - Model size (float): 4MB
61
+ - Number of parameters: 1M
62
+
63
+ ## Performance Summary
64
+ | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
65
+ |---|---|---|---|---|---|---
66
+ | MobileFaceNet | ONNX | float | Snapdragon® X2 Elite | 0.583 ms | 1 - 1 MB | NPU
67
+ | MobileFaceNet | ONNX | float | Snapdragon® X Elite | 1.038 ms | 0 - 0 MB | NPU
68
+ | MobileFaceNet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.7 ms | 0 - 51 MB | NPU
69
+ | MobileFaceNet | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 1.511 ms | 0 - 54 MB | NPU
70
+ | MobileFaceNet | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 1.471 ms | 0 - 4 MB | NPU
71
+ | MobileFaceNet | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 1.002 ms | 0 - 4 MB | NPU
72
+ | MobileFaceNet | ONNX | float | Qualcomm® QCS8450 | 1.511 ms | 0 - 54 MB | NPU
73
+ | MobileFaceNet | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 1.38 ms | 0 - 3 MB | NPU
74
+ | MobileFaceNet | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 1.038 ms | 0 - 0 MB | NPU
75
+ | MobileFaceNet | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 0.567 ms | 0 - 36 MB | NPU
76
+ | MobileFaceNet | ONNX | float | Snapdragon® 8 Elite Mobile | 0.567 ms | 0 - 36 MB | NPU
77
+ | MobileFaceNet | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.51 ms | 0 - 33 MB | NPU
78
+ | MobileFaceNet | ONNX | w8a16 | Snapdragon® X2 Elite | 0.446 ms | 1 - 1 MB | NPU
79
+ | MobileFaceNet | ONNX | w8a16 | Snapdragon® X Elite | 0.824 ms | 0 - 0 MB | NPU
80
+ | MobileFaceNet | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 0.581 ms | 0 - 53 MB | NPU
81
+ | MobileFaceNet | ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 1.103 ms | 0 - 57 MB | NPU
82
+ | MobileFaceNet | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 2.958 ms | 0 - 3 MB | NPU
83
+ | MobileFaceNet | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 0.851 ms | 0 - 4 MB | NPU
84
+ | MobileFaceNet | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.804 ms | 0 - 4 MB | NPU
85
+ | MobileFaceNet | ONNX | w8a16 | Qualcomm® QCS8450 | 1.103 ms | 0 - 57 MB | NPU
86
+ | MobileFaceNet | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 0.914 ms | 0 - 3 MB | NPU
87
+ | MobileFaceNet | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 0.824 ms | 0 - 0 MB | NPU
88
+ | MobileFaceNet | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 5.426 ms | 0 - 161 MB | NPU
89
+ | MobileFaceNet | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 0.909 ms | 0 - 41 MB | NPU
90
+ | MobileFaceNet | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 0.459 ms | 0 - 40 MB | NPU
91
+ | MobileFaceNet | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 0.459 ms | 0 - 40 MB | NPU
92
+ | MobileFaceNet | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 0.369 ms | 0 - 42 MB | NPU
93
+ | MobileFaceNet | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 0.909 ms | 0 - 41 MB | NPU
94
+ | MobileFaceNet | QNN_DLC | float | Snapdragon® X2 Elite | 0.696 ms | 0 - 0 MB | NPU
95
+ | MobileFaceNet | QNN_DLC | float | Snapdragon® X Elite | 1.305 ms | 0 - 0 MB | NPU
96
+ | MobileFaceNet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 0.807 ms | 0 - 46 MB | NPU
97
+ | MobileFaceNet | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 1.63 ms | 0 - 50 MB | NPU
98
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 1.467 ms | 0 - 3 MB | NPU
99
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 4.53 ms | 0 - 30 MB | NPU
100
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 1.15 ms | 0 - 12 MB | NPU
101
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® SA8775P | 1.674 ms | 0 - 33 MB | NPU
102
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® SA8650P | 1.674 ms | 0 - 33 MB | NPU
103
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® SA8255P | 1.674 ms | 0 - 33 MB | NPU
104
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® QCS8450 | 1.63 ms | 0 - 50 MB | NPU
105
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 1.499 ms | 0 - 2 MB | NPU
106
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 1.305 ms | 0 - 0 MB | NPU
107
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 0.606 ms | 0 - 30 MB | NPU
108
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® SA7255P | 4.53 ms | 0 - 30 MB | NPU
109
+ | MobileFaceNet | QNN_DLC | float | Qualcomm® SA8295P | 2.034 ms | 0 - 29 MB | NPU
110
+ | MobileFaceNet | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 0.606 ms | 0 - 30 MB | NPU
111
+ | MobileFaceNet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.478 ms | 0 - 31 MB | NPU
112
+ | MobileFaceNet | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 0.605 ms | 0 - 0 MB | NPU
113
+ | MobileFaceNet | QNN_DLC | w8a16 | Snapdragon® X Elite | 1.156 ms | 0 - 0 MB | NPU
114
+ | MobileFaceNet | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 0.697 ms | 0 - 47 MB | NPU
115
+ | MobileFaceNet | QNN_DLC | w8a16 | Snapdragon® 8 Gen 1 Mobile | 1.252 ms | 0 - 48 MB | NPU
116
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 3.611 ms | 0 - 2 MB | NPU
117
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 0.992 ms | 0 - 2 MB | NPU
118
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 2.217 ms | 0 - 36 MB | NPU
119
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.995 ms | 0 - 71 MB | NPU
120
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® SA8775P | 1.198 ms | 0 - 38 MB | NPU
121
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® SA8650P | 1.198 ms | 0 - 38 MB | NPU
122
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® SA8255P | 1.198 ms | 0 - 38 MB | NPU
123
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® QCS8450 | 1.252 ms | 0 - 48 MB | NPU
124
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 1.095 ms | 0 - 2 MB | NPU
125
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 1.156 ms | 0 - 0 MB | NPU
126
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 6.123 ms | 0 - 149 MB | NPU
127
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 1.104 ms | 0 - 36 MB | NPU
128
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 0.489 ms | 0 - 35 MB | NPU
129
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® SA7255P | 2.217 ms | 0 - 36 MB | NPU
130
+ | MobileFaceNet | QNN_DLC | w8a16 | Qualcomm® SA8295P | 1.584 ms | 0 - 34 MB | NPU
131
+ | MobileFaceNet | QNN_DLC | w8a16 | Snapdragon® 8 Elite Mobile | 0.489 ms | 0 - 35 MB | NPU
132
+ | MobileFaceNet | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 0.398 ms | 0 - 37 MB | NPU
133
+ | MobileFaceNet | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 1.104 ms | 0 - 36 MB | NPU
134
+ | MobileFaceNet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 0.687 ms | 0 - 48 MB | NPU
135
+ | MobileFaceNet | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 1.481 ms | 0 - 49 MB | NPU
136
+ | MobileFaceNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 1.444 ms | 0 - 6 MB | NPU
137
+ | MobileFaceNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 4.498 ms | 0 - 33 MB | NPU
138
+ | MobileFaceNet | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.977 ms | 0 - 2 MB | NPU
139
+ | MobileFaceNet | TFLITE | float | Qualcomm® SA8775P | 1.502 ms | 0 - 36 MB | NPU
140
+ | MobileFaceNet | TFLITE | float | Qualcomm® SA8650P | 1.502 ms | 0 - 36 MB | NPU
141
+ | MobileFaceNet | TFLITE | float | Qualcomm® SA8255P | 1.502 ms | 0 - 36 MB | NPU
142
+ | MobileFaceNet | TFLITE | float | Qualcomm® QCS8450 | 1.481 ms | 0 - 49 MB | NPU
143
+ | MobileFaceNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 1.347 ms | 0 - 5 MB | NPU
144
+ | MobileFaceNet | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 0.575 ms | 0 - 32 MB | NPU
145
+ | MobileFaceNet | TFLITE | float | Qualcomm® SA7255P | 4.498 ms | 0 - 33 MB | NPU
146
+ | MobileFaceNet | TFLITE | float | Qualcomm® SA8295P | 1.81 ms | 0 - 28 MB | NPU
147
+ | MobileFaceNet | TFLITE | float | Snapdragon® 8 Elite Mobile | 0.575 ms | 0 - 32 MB | NPU
148
+ | MobileFaceNet | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.509 ms | 0 - 31 MB | NPU
149
+
150
+ ## License
151
+ * The license for the original implementation of MobileFaceNet can be found
152
+ [here](https://github.com/foamliu/MobileFaceNet/blob/master/LICENSE).
153
+
154
+ ## References
155
+ * [Source Model Implementation](https://github.com/foamliu/MobileFaceNet)
156
+
157
+ ## Community
158
+ * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
159
+ * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).
release_assets.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "0.61.0",
3
+ "precisions": {
4
+ "float": {
5
+ "universal_assets": {
6
+ "onnx": {
7
+ "tool_versions": {
8
+ "qairt": "2.45.0.260326154327",
9
+ "onnx_runtime": "1.27.1"
10
+ },
11
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/mobile_facenet/releases/v0.61.0/mobile_facenet-onnx-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/mobile_facenet/releases/v0.61.0/mobile_facenet-qnn_dlc-float.zip"
18
+ },
19
+ "tflite": {
20
+ "tool_versions": {
21
+ "qairt": "2.45.0.260326154327"
22
+ },
23
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/mobile_facenet/releases/v0.61.0/mobile_facenet-tflite-float.zip"
24
+ }
25
+ }
26
+ },
27
+ "w8a16": {
28
+ "universal_assets": {
29
+ "onnx": {
30
+ "tool_versions": {
31
+ "qairt": "2.45.0.260326154327",
32
+ "onnx_runtime": "1.27.1"
33
+ },
34
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/mobile_facenet/releases/v0.61.0/mobile_facenet-onnx-w8a16.zip"
35
+ },
36
+ "qnn_dlc": {
37
+ "tool_versions": {
38
+ "qairt": "2.45.0.260326154327"
39
+ },
40
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/mobile_facenet/releases/v0.61.0/mobile_facenet-qnn_dlc-w8a16.zip"
41
+ }
42
+ }
43
+ }
44
+ }
45
+ }