Hugging Face's logo

qualcomm
/
RegNet

Image Classification
PyTorch
backbone
android
Model card Files
xet
 Community
RegNet
File size: 9,805 Bytes
a97ffba
 
0e88ee3
a97ffba
 
 
abcd9fe
a97ffba
 
 
8c697f6
a97ffba
8c067c7
ea567a9
a97ffba
 
8c067c7
df4d6a4
8c067c7
 
 
 
 
 
 
 
 
 
 
 
31c5b18
 
 
 
 
 
8c067c7
 
 
 
 
 
df4d6a4
8c067c7
 
 
 
 
 
df4d6a4
8c067c7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31c5b18
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c20270d
a97ffba
4489a96
 
c20270d
a97ffba
 
 
 
 
82e0e91
a97ffba
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:
- backbone
- android
pipeline_tag: image-classification

---

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

# RegNet: Optimized for Qualcomm Devices

RegNet is a machine learning model that can classify images from the Imagenet dataset. It can also be used as a backbone in building more complex models for specific use cases.

This is based on the implementation of RegNet found [here](https://github.com/pytorch/vision/blob/main/torchvision/models/regnet.py).
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/regnet) 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/regnet/releases/v0.50.2/regnet-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/regnet/releases/v0.50.2/regnet-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/regnet/releases/v0.50.2/regnet-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/regnet/releases/v0.50.2/regnet-qnn_dlc-w8a8.zip)
| TFLITE | float | Universal | QAIRT 2.43, TFLite 2.19.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/regnet/releases/v0.50.2/regnet-tflite-float.zip)
| TFLITE | w8a8 | Universal | QAIRT 2.43, TFLite 2.19.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/regnet/releases/v0.50.2/regnet-tflite-w8a8.zip)

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


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

## Model Details

**Model Type:** Model_use_case.image_classification

**Model Stats:**
- Model checkpoint: Imagenet
- Input resolution: 224x224
- Number of parameters: 15.3M
- Model size (float): 58.3 MB
- Model size (w8a8): 15.4 MB

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| RegNet | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.832 ms | 1 - 88 MB | NPU
| RegNet | ONNX | float | Snapdragon® X2 Elite | 0.905 ms | 39 - 39 MB | NPU
| RegNet | ONNX | float | Snapdragon® X Elite | 1.979 ms | 39 - 39 MB | NPU
| RegNet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 1.244 ms | 0 - 135 MB | NPU
| RegNet | ONNX | float | Qualcomm® QCS8550 (Proxy) | 1.754 ms | 1 - 4 MB | NPU
| RegNet | ONNX | float | Qualcomm® QCS9075 | 2.856 ms | 0 - 4 MB | NPU
| RegNet | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.997 ms | 0 - 87 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.505 ms | 0 - 89 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® X2 Elite | 0.483 ms | 20 - 20 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® X Elite | 1.116 ms | 20 - 20 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.647 ms | 0 - 135 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® QCS6490 | 27.892 ms | 7 - 19 MB | CPU
| RegNet | ONNX | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.902 ms | 0 - 28 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® QCS9075 | 1.079 ms | 0 - 3 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® QCM6690 | 18.24 ms | 9 - 18 MB | CPU
| RegNet | ONNX | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.547 ms | 0 - 79 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 13.774 ms | 9 - 18 MB | CPU
| RegNet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.896 ms | 1 - 81 MB | NPU
| RegNet | QNN_DLC | float | Snapdragon® X2 Elite | 1.272 ms | 1 - 1 MB | NPU
| RegNet | QNN_DLC | float | Snapdragon® X Elite | 2.313 ms | 1 - 1 MB | NPU
| RegNet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 1.39 ms | 0 - 128 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 9.913 ms | 1 - 77 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 2.087 ms | 1 - 2 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® SA8775P | 3.316 ms | 0 - 78 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® QCS9075 | 3.064 ms | 1 - 3 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 3.473 ms | 0 - 115 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® SA7255P | 9.913 ms | 1 - 77 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® SA8295P | 3.513 ms | 0 - 65 MB | NPU
| RegNet | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.142 ms | 0 - 79 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.445 ms | 0 - 80 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 0.592 ms | 0 - 0 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® X Elite | 1.064 ms | 0 - 0 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.631 ms | 0 - 107 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® QCS6490 | 2.701 ms | 0 - 2 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® QCS8275 (Proxy) | 2.288 ms | 0 - 76 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.881 ms | 0 - 88 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® SA8775P | 1.299 ms | 0 - 78 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® QCS9075 | 1.081 ms | 0 - 2 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® QCM6690 | 7.096 ms | 0 - 196 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® QCS8450 (Proxy) | 1.299 ms | 0 - 111 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® SA7255P | 2.288 ms | 0 - 76 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® SA8295P | 1.632 ms | 0 - 75 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.488 ms | 0 - 79 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 1.173 ms | 0 - 77 MB | NPU
| RegNet | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.899 ms | 0 - 104 MB | NPU
| RegNet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 1.397 ms | 0 - 159 MB | NPU
| RegNet | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 9.9 ms | 0 - 101 MB | NPU
| RegNet | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 2.068 ms | 0 - 2 MB | NPU
| RegNet | TFLITE | float | Qualcomm® SA8775P | 3.342 ms | 0 - 102 MB | NPU
| RegNet | TFLITE | float | Qualcomm® QCS9075 | 3.053 ms | 0 - 42 MB | NPU
| RegNet | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 3.449 ms | 0 - 143 MB | NPU
| RegNet | TFLITE | float | Qualcomm® SA7255P | 9.9 ms | 0 - 101 MB | NPU
| RegNet | TFLITE | float | Qualcomm® SA8295P | 3.48 ms | 0 - 83 MB | NPU
| RegNet | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.107 ms | 0 - 95 MB | NPU
| RegNet | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 Mobile | 0.389 ms | 0 - 79 MB | NPU
| RegNet | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.523 ms | 0 - 118 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® QCS6490 | 2.374 ms | 0 - 21 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® QCS8275 (Proxy) | 1.978 ms | 0 - 76 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® QCS8550 (Proxy) | 0.727 ms | 0 - 126 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® SA8775P | 1.144 ms | 0 - 77 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® QCS9075 | 0.893 ms | 0 - 22 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® QCM6690 | 6.606 ms | 0 - 191 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® QCS8450 (Proxy) | 1.097 ms | 0 - 111 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® SA7255P | 1.978 ms | 0 - 76 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® SA8295P | 1.419 ms | 0 - 72 MB | NPU
| RegNet | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.424 ms | 0 - 70 MB | NPU
| RegNet | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.987 ms | 0 - 73 MB | NPU

## License
* The license for the original implementation of RegNet can be found
  [here](https://github.com/pytorch/vision/blob/main/LICENSE).

## References
* [Designing Network Design Spaces](https://arxiv.org/abs/2003.13678)
* [Source Model Implementation](https://github.com/pytorch/vision/blob/main/torchvision/models/regnet.py)

## 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).