Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/sicara/tf2-yolov4

A TensorFlow 2.0 implementation of YOLOv4: Optimal Speed and Accuracy of Object Detection
https://github.com/sicara/tf2-yolov4

keras object-detection tensorflow yolo yolov4

Last synced: 3 days ago
JSON representation

A TensorFlow 2.0 implementation of YOLOv4: Optimal Speed and Accuracy of Object Detection

Host: GitHub
URL: https://github.com/sicara/tf2-yolov4
Owner: sicara
License: mit
Created: 2020-04-28T07:06:28.000Z (almost 5 years ago)
Default Branch: master
Last Pushed: 2024-03-20T15:33:12.000Z (11 months ago)
Last Synced: 2025-02-09T21:07:18.191Z (10 days ago)
Topics: keras, object-detection, tensorflow, yolo, yolov4
Language: Python
Size: 3.26 MB
Stars: 208
Watchers: 14
Forks: 53
Open Issues: 24
Metadata Files:
- Readme: README.md
- License: LICENSE

Awesome Lists containing this project

awesome-yolo-object-detection - sicara/tf2-yolov4 - yolov4?style=social"/> : A TensorFlow 2.0 implementation of YOLOv4: Optimal Speed and Accuracy of Object Detection. (Other Versions of YOLO)
awesome-yolo-object-detection - sicara/tf2-yolov4 - yolov4?style=social"/> : A TensorFlow 2.0 implementation of YOLOv4: Optimal Speed and Accuracy of Object Detection. (Other Versions of YOLO)

README

        # YOLOv4

> A TensorFlow 2.0 implementation of YOLOv4: Optimal Speed and Accuracy of Object Detection

[![Pypi Version](https://img.shields.io/pypi/v/tf2-yolov4.svg)](https://pypi.org/project/tf2-yolov4/)

![Python Versions](https://img.shields.io/badge/python-3.6%20|%203.7%20|%203.8-%23EBBD68.svg)

![Tensorflow Versions](https://img.shields.io/badge/TensorFlow-2.x-blue.svg)

[![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/python/black)

[![Python package](https://github.com/sicara/tf2-yolov4/workflows/Python%20package/badge.svg?branch=master)](https://github.com/sicara/tf2-yolov4/actions?query=workflow%3A%22Python+package%22)

[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/sicara/tf2-yolov4/blob/master/notebooks/YoloV4_Dectection_Example.ipynb)

This implementation runs (for now) inference with the original Darknet weights from [AlexeyAB](https://www.github.com/AlexeyAB/darknet).

See the roadmap section to see what's next.



    



## Installation

To install this package, you can run:

```bash

pip install tf2_yolov4

pip install tensorflow

# Check that tf2_yolov4 is installed properly

python -c "from tf2_yolov4.model import YOLOv4; print(YOLOv4)"

```

Requirements:

- MacOs >= 10.15 since tensorflow-addons is not available for older release of MacOs

- Python >= 3.6

- Compatible versions between TensorFlow and TensorFlow Addons: check the [compatibility matrix](https://github.com/tensorflow/addons#python-op-compatibility-matrix)

## Examples in Colab

- [Run detection on a single image](./notebooks/YoloV4_Dectection_Example.ipynb) / [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/sicara/tf2-yolov4/blob/master/notebooks/YoloV4_Dectection_Example.ipynb)

## Pretrained weights

Our YOLOv4 implementation supports the `weights` argument similarly to Keras applications. To load a model with pretrained

weights, you can simply call:

```python

# Loads Darknet weights trained on COCO

model = YOLOv4(

    input_shape,

    num_classes,

    anchors,

    weights="darknet",

)

```

If weights are available locally, they will be used. Otherwise, they will be automatically downloaded.

## Roadmap

- [x] Inference

    - [x] CSPDarknet53 backbone with Mish activations

    - [x] SPP Neck

    - [x] YOLOv3 Head

    - [x] Load Darknet Weights

    - [x] Image loading and preprocessing

    - [x] YOLOv3 box postprocessing

    - [x] Handling non-square images

- [ ] Training

    - [ ] Training loop with YOLOv3 loss

    - [ ] CIoU loss

    - [ ] Cross mini-Batch Normalization

    - [ ] Self-adversarial Training

    - [ ] Mosaic Data Augmentation

    - [ ] DropBlock

- [ ] Enhancements

    - [x] Automatic download of pretrained weights (like Keras applications)

## References

- [yolov3-tf2](https://github.com/zzh8829/yolov3-tf2)