Hugging Face's logo

qualcomm
/
ConvNext-Base

Image Classification
PyTorch
bu_auto
android
Model card Files
xet
 Community
ConvNext-Base
File size: 11,858 Bytes
983311f
 
e6cb09f
983311f
9dfb2ad
983311f
e9c2552
983311f
 
 
 
 
e1a78fd
983311f
 
 
e1a78fd
2db064f
e1a78fd
 
 
 
 
 
 
 
 
 
 
 
2db064f
 
 
 
 
e1a78fd
 
 
 
 
 
2db064f
e1a78fd
 
 
 
 
 
2db064f
e1a78fd
 
 
 
 
 
 
2db064f
e1a78fd
 
2db064f
e1a78fd
 
 
 
2db064f
 
 
 
 
 
 
 
 
 
 
cc54a19
2db064f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
983311f
 
7f2d655
 
983311f
 
 
 
 
 
 
 
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
150
151
152
153
154
155
156
157
158
159
---
library_name: pytorch
license: other
tags:
- bu_auto
- android
pipeline_tag: image-classification

---

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

# ConvNext-Base: Optimized for Qualcomm Devices

ConvNextBase 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 ConvNext-Base found [here](https://github.com/pytorch/vision/blob/main/torchvision/models/convnext.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/v0.60.0/src/qai_hub_models/models/convnext_base) 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 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/convnext_base/releases/v0.60.0/convnext_base-onnx-float.zip)
| 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/convnext_base/releases/v0.60.0/convnext_base-onnx-w8a16.zip)
| QNN_DLC | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/convnext_base/releases/v0.60.0/convnext_base-qnn_dlc-float.zip)
| QNN_DLC | w8a16 | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/convnext_base/releases/v0.60.0/convnext_base-qnn_dlc-w8a16.zip)
| TFLITE | float | Universal | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/convnext_base/releases/v0.60.0/convnext_base-tflite-float.zip)

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


### Option 2: Export with Custom Configurations

Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.60.0/src/qai_hub_models/models/convnext_base) 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 [ConvNext-Base on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.60.0/src/qai_hub_models/models/convnext_base) for usage instructions.

## Model Details

**Model Type:** Model_use_case.image_classification

**Model Stats:**
- Input resolution: 224x224
- Model checkpoint: Imagenet
- Model size (float): 338 MB
- Model size (w8a16): 88.7 MB
- Number of parameters: 88.6M

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| ConvNext-Base | ONNX | float | Snapdragon® X2 Elite | 3.504 ms | 2 - 2 MB | NPU
| ConvNext-Base | ONNX | float | Snapdragon® X Elite | 7.232 ms | 176 - 176 MB | NPU
| ConvNext-Base | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 5.296 ms | 1 - 313 MB | NPU
| ConvNext-Base | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 19.215 ms | 1 - 301 MB | NPU
| ConvNext-Base | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 7.134 ms | 0 - 245 MB | NPU
| ConvNext-Base | ONNX | float | Qualcomm® QCS8450 | 19.215 ms | 1 - 301 MB | NPU
| ConvNext-Base | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 10.831 ms | 0 - 4 MB | NPU
| ConvNext-Base | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 7.232 ms | 176 - 176 MB | NPU
| ConvNext-Base | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 4.128 ms | 0 - 185 MB | NPU
| ConvNext-Base | ONNX | float | Snapdragon® 8 Elite Mobile | 4.128 ms | 0 - 185 MB | NPU
| ConvNext-Base | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.201 ms | 0 - 186 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Snapdragon® X2 Elite | 2.383 ms | 1 - 1 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Snapdragon® X Elite | 4.977 ms | 91 - 91 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.424 ms | 0 - 262 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 8.063 ms | 0 - 260 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 33.134 ms | 0 - 3 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 4.901 ms | 0 - 4 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 4.867 ms | 0 - 4 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Qualcomm® QCS8450 | 8.063 ms | 0 - 260 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 4.883 ms | 0 - 3 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 4.977 ms | 91 - 91 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 58.166 ms | 1 - 400 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 6.881 ms | 0 - 261 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 2.778 ms | 0 - 204 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 2.778 ms | 0 - 204 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.173 ms | 0 - 225 MB | NPU
| ConvNext-Base | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 6.881 ms | 0 - 261 MB | NPU
| ConvNext-Base | QNN_DLC | float | Snapdragon® X2 Elite | 4.233 ms | 1 - 1 MB | NPU
| ConvNext-Base | QNN_DLC | float | Snapdragon® X Elite | 8.41 ms | 1 - 1 MB | NPU
| ConvNext-Base | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 5.874 ms | 0 - 307 MB | NPU
| ConvNext-Base | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 20.393 ms | 0 - 296 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 12.775 ms | 1 - 4 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 41.893 ms | 1 - 180 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 7.982 ms | 1 - 2 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® SA8775P | 11.795 ms | 1 - 181 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® SA8650P | 11.795 ms | 1 - 181 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® SA8255P | 11.795 ms | 1 - 181 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® QCS8450 | 20.393 ms | 0 - 296 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 11.653 ms | 1 - 3 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 8.41 ms | 1 - 1 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 4.586 ms | 1 - 183 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® SA7255P | 41.893 ms | 1 - 180 MB | NPU
| ConvNext-Base | QNN_DLC | float | Qualcomm® SA8295P | 19.676 ms | 1 - 171 MB | NPU
| ConvNext-Base | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 4.586 ms | 1 - 183 MB | NPU
| ConvNext-Base | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.55 ms | 1 - 186 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 3.096 ms | 0 - 0 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Snapdragon® X Elite | 6.329 ms | 0 - 0 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 4.074 ms | 0 - 250 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Snapdragon® 8 Gen 1 Mobile | 9.074 ms | 0 - 250 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 29.883 ms | 0 - 2 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 5.768 ms | 0 - 3 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 14.627 ms | 0 - 204 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.9 ms | 0 - 2 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® SA8775P | 6.2 ms | 0 - 205 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® SA8650P | 6.2 ms | 0 - 205 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® SA8255P | 6.2 ms | 0 - 205 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® QCS8450 | 9.074 ms | 0 - 250 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 5.931 ms | 2 - 4 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 6.329 ms | 0 - 0 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 75.164 ms | 0 - 402 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 7.742 ms | 0 - 256 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 3.256 ms | 0 - 195 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® SA7255P | 14.627 ms | 0 - 204 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Qualcomm® SA8295P | 9.386 ms | 0 - 204 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Snapdragon® 8 Elite Mobile | 3.256 ms | 0 - 195 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.573 ms | 0 - 214 MB | NPU
| ConvNext-Base | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 7.742 ms | 0 - 256 MB | NPU
| ConvNext-Base | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 5.449 ms | 0 - 302 MB | NPU
| ConvNext-Base | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 19.656 ms | 0 - 292 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 12.179 ms | 0 - 178 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 40.981 ms | 0 - 175 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 7.311 ms | 0 - 3 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® SA8775P | 11.074 ms | 0 - 176 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® SA8650P | 11.074 ms | 0 - 176 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® SA8255P | 11.074 ms | 0 - 176 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® QCS8450 | 19.656 ms | 0 - 292 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 10.811 ms | 0 - 177 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 4.091 ms | 0 - 179 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® SA7255P | 40.981 ms | 0 - 175 MB | NPU
| ConvNext-Base | TFLITE | float | Qualcomm® SA8295P | 18.78 ms | 0 - 161 MB | NPU
| ConvNext-Base | TFLITE | float | Snapdragon® 8 Elite Mobile | 4.091 ms | 0 - 179 MB | NPU
| ConvNext-Base | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 3.219 ms | 0 - 180 MB | NPU

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

## References
* [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545)
* [Source Model Implementation](https://github.com/pytorch/vision/blob/main/torchvision/models/convnext.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).