---
library_name: pytorch
license: bsd-3-clause
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/v0.63.0/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 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/regnet/releases/v0.63.0/regnet-onnx-float.zip)
| ONNX | w8a8 | Universal | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/regnet/releases/v0.63.0/regnet-onnx-w8a8.zip)
| QNN_DLC | float | Universal | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/regnet/releases/v0.63.0/regnet-qnn_dlc-float.zip)
| QNN_DLC | w8a8 | Universal | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/regnet/releases/v0.63.0/regnet-qnn_dlc-w8a8.zip)
| TFLITE | float | Universal | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/regnet/releases/v0.63.0/regnet-tflite-float.zip)
| TFLITE | w8a8 | Universal | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/regnet/releases/v0.63.0/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/v0.63.0/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/v0.63.0/src/qai_hub_models/models/regnet) for usage instructions.

## Model Details

**Model Type:** Model_use_case.image_classification

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

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| RegNet | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.832 ms | 0 - 80 MB | NPU
| RegNet | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.989 ms | 0 - 80 MB | NPU
| RegNet | ONNX | float | Snapdragon® X2 Elite | 0.891 ms | 2 - 2 MB | NPU
| RegNet | ONNX | float | Snapdragon® X Elite | 1.763 ms | 39 - 39 MB | NPU
| RegNet | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 1.202 ms | 0 - 131 MB | NPU
| RegNet | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 2.576 ms | 0 - 121 MB | NPU
| RegNet | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 3.026 ms | 0 - 5 MB | NPU
| RegNet | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 1.736 ms | 0 - 41 MB | NPU
| RegNet | ONNX | float | Qualcomm® QCS8450 | 2.576 ms | 0 - 121 MB | NPU
| RegNet | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 2.81 ms | 0 - 4 MB | NPU
| RegNet | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 1.763 ms | 39 - 39 MB | NPU
| RegNet | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 0.989 ms | 0 - 80 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.517 ms | 0 - 94 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.553 ms | 0 - 93 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® X2 Elite | 0.435 ms | 1 - 1 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® X Elite | 0.923 ms | 21 - 21 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.652 ms | 0 - 142 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® 8 Gen 1 Mobile | 1.117 ms | 0 - 138 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 2.629 ms | 0 - 3 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® Dragonwing™ IQ-8275 | 1.021 ms | 0 - 4 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.893 ms | 0 - 26 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® QCS8450 | 1.117 ms | 0 - 138 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 1.154 ms | 0 - 3 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® Dragonwing™ IQ-X7181 | 0.923 ms | 21 - 21 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 7.129 ms | 0 - 201 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 1.094 ms | 0 - 90 MB | NPU
| RegNet | ONNX | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 0.553 ms | 0 - 93 MB | NPU
| RegNet | ONNX | w8a8 | Snapdragon® 7 Gen 4 Mobile | 1.094 ms | 0 - 90 MB | NPU
| RegNet | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.906 ms | 1 - 81 MB | NPU
| RegNet | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.101 ms | 0 - 77 MB | NPU
| RegNet | QNN_DLC | float | Snapdragon® X2 Elite | 1.234 ms | 1 - 1 MB | NPU
| RegNet | QNN_DLC | float | Snapdragon® X Elite | 2.254 ms | 1 - 1 MB | NPU
| RegNet | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 1.404 ms | 0 - 128 MB | NPU
| RegNet | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 3.352 ms | 0 - 122 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 3.178 ms | 1 - 4 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 2.043 ms | 1 - 2 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® QCS8450 | 3.352 ms | 0 - 122 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 3.024 ms | 1 - 3 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 2.254 ms | 1 - 1 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 1.101 ms | 0 - 77 MB | NPU
| RegNet | QNN_DLC | float | Qualcomm® SA8295P | 3.336 ms | 0 - 65 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.444 ms | 0 - 85 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.479 ms | 0 - 87 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® X2 Elite | 0.537 ms | 0 - 0 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® X Elite | 1.044 ms | 0 - 0 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.625 ms | 0 - 116 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® 8 Gen 1 Mobile | 1.099 ms | 0 - 115 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 2.679 ms | 2 - 4 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-8275 | 0.948 ms | 0 - 3 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.87 ms | 0 - 2 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® QCS8450 | 1.099 ms | 0 - 115 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 1.107 ms | 0 - 2 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ IQ-X7181 | 1.044 ms | 0 - 0 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 7.436 ms | 0 - 195 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 1.166 ms | 0 - 83 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 0.479 ms | 0 - 87 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® SA7255P | 2.221 ms | 0 - 84 MB | NPU
| RegNet | QNN_DLC | w8a8 | Qualcomm® SA8295P | 1.463 ms | 0 - 82 MB | NPU
| RegNet | QNN_DLC | w8a8 | Snapdragon® 7 Gen 4 Mobile | 1.166 ms | 0 - 83 MB | NPU
| RegNet | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.901 ms | 0 - 99 MB | NPU
| RegNet | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 1.094 ms | 0 - 99 MB | NPU
| RegNet | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 1.39 ms | 4 - 158 MB | NPU
| RegNet | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 3.294 ms | 0 - 144 MB | NPU
| RegNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 3.17 ms | 0 - 43 MB | NPU
| RegNet | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 2.018 ms | 0 - 2 MB | NPU
| RegNet | TFLITE | float | Qualcomm® SA8775P | 14.662 ms | 0 - 28 MB | GPU
| RegNet | TFLITE | float | Qualcomm® SA8650P | 14.662 ms | 0 - 28 MB | GPU
| RegNet | TFLITE | float | Qualcomm® SA8255P | 14.662 ms | 0 - 28 MB | GPU
| RegNet | TFLITE | float | Qualcomm® QCS8450 | 3.294 ms | 0 - 144 MB | NPU
| RegNet | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 3.031 ms | 0 - 42 MB | NPU
| RegNet | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 1.094 ms | 0 - 99 MB | NPU
| RegNet | TFLITE | float | Qualcomm® SA7255P | 9.849 ms | 0 - 93 MB | NPU
| RegNet | TFLITE | float | Qualcomm® SA8295P | 3.324 ms | 0 - 77 MB | NPU
| RegNet | TFLITE | w8a8 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.392 ms | 0 - 76 MB | NPU
| RegNet | TFLITE | w8a8 | Snapdragon® 8 Elite For Galaxy Mobile | 0.432 ms | 0 - 83 MB | NPU
| RegNet | TFLITE | w8a8 | Snapdragon® 8 Gen 3 Mobile | 0.507 ms | 0 - 114 MB | NPU
| RegNet | TFLITE | w8a8 | Snapdragon® 8 Gen 1 Mobile | 0.922 ms | 0 - 120 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS6490 | 2.43 ms | 0 - 22 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® Dragonwing™ IQ-8275 | 0.809 ms | 0 - 23 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.703 ms | 0 - 2 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® SA8775P | 1.068 ms | 0 - 82 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® SA8650P | 1.068 ms | 0 - 82 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® SA8255P | 1.068 ms | 0 - 82 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® QCS8450 | 0.922 ms | 0 - 120 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® Dragonwing™ IQ-9075 | 0.916 ms | 0 - 22 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-6690 | 6.533 ms | 0 - 188 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-7790 | 0.979 ms | 0 - 77 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® Dragonwing™ Q-8750 | 0.432 ms | 0 - 83 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® SA7255P | 1.92 ms | 0 - 79 MB | NPU
| RegNet | TFLITE | w8a8 | Qualcomm® SA8295P | 1.275 ms | 0 - 75 MB | NPU
| RegNet | TFLITE | w8a8 | Snapdragon® 7 Gen 4 Mobile | 0.979 ms | 0 - 77 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).