https://github.com/Hyuto/yolo-nas-onnx

Inference YOLO-NAS ONNX model
https://github.com/Hyuto/yolo-nas-onnx

cpp javascript object-detection onnx onnxruntime opencv opencv-dnn python yolo-nas

Last synced: 3 months ago
JSON representation

Inference YOLO-NAS ONNX model

• Host: GitHub
• URL: https://github.com/Hyuto/yolo-nas-onnx
• Owner: Hyuto
• License: mit
• Created: 2023-05-11T17:12:12.000Z (almost 2 years ago)
• Default Branch: master
• Last Pushed: 2023-12-18T07:43:06.000Z (about 1 year ago)
• Last Synced: 2024-08-02T01:19:41.072Z (6 months ago)
• Topics: cpp, javascript, object-detection, onnx, onnxruntime, opencv, opencv-dnn, python, yolo-nas
• Language: C++
• Homepage: https://hyuto.github.io/yolo-nas-onnx/
• Size: 65.7 MB
• Stars: 60
• Watchers: 3
• Forks: 6
• Open Issues: 3
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-yolo-object-detection - Hyuto/yolo-nas-onnx - nas-onnx?style=social"/> : Inference YOLO-NAS ONNX model. [hyuto.github.io/yolo-nas-onnx/](https://hyuto.github.io/yolo-nas-onnx/) (Lighter and Deployment Frameworks)
• awesome-yolo-object-detection - Hyuto/yolo-nas-onnx - nas-onnx?style=social"/> : Inference YOLO-NAS ONNX model. [hyuto.github.io/yolo-nas-onnx/](https://hyuto.github.io/yolo-nas-onnx/) (Lighter and Deployment Frameworks)

README

        
# YOLO-NAS ONNX



    



**_Image Source_**: https://www.pinterest.com/pin/784752303797219490/

---

![love](https://img.shields.io/badge/Made%20with-🖤-white)

![onnxruntime-web](https://img.shields.io/badge/onnxruntime--web-white?logo=onnx&logoColor=black)

![opencv](https://img.shields.io/badge/OpenCV-4.7.0-white?logo=opencv)

![python](https://img.shields.io/badge/Python-darkgreen?logo=python)

![c++](https://img.shields.io/badge/C++-red?logo=cplusplus)

![javascript](https://img.shields.io/badge/JavaScript-green?logo=javascript)

Run YOLO-NAS models with ONNX **without using Pytorch**. Inferencing YOLO-NAS ONNX models with ONNXRUNTIME or OpenCV DNN.

## Generate ONNX Model

Generate YOLO-NAS ONNX model **without preprocessing and postprocessing within the model**.

You can convert the model using the following code after installing `super_gradients` library.

**Example: Exporting YOLO-NAS S**

```python

from super_gradients.training import models

from super_gradients.common.object_names import Models

model = models.get(Models.YOLO_NAS_S, pretrained_weights="coco")

model.eval()

model.prep_model_for_conversion(input_size=[1, 3, 640, 640])

model.export("yolo_nas_s.onnx", postprocessing=None, preprocessing=None)

```

## Custom Model

To run custom trained YOLO-NAS model in this project you need to generate custom model metadata.

Custom model metadata generated from [custom-nas-model-metadata.py](https://gist.github.com/Hyuto/f3db1c0c2c36308284e101f441c2555f)

to provide additional information from torch model.

**Usage**

```bash

python custom-nas-model-metadata.py -m  \ # Custom trained YOLO-NAS checkpoint path

                                        -t  \ # Custom trained YOLO-NAS model type

                                        -n  # Number of classes

```

After running that it'll generate metadata (json formated) for you

## References

- https://github.com/Deci-AI/super-gradients