Skip to content

Latest commit

 

History

History
68 lines (51 loc) · 2.32 KB

README.md

File metadata and controls

68 lines (51 loc) · 2.32 KB
Raw

YOLOv5 Trainer

The YOLOv5 Trainer is a Python program designed to streamline the process of training YOLOv5 models on custom datasets. It provides a easy way to model training, testing, and model management.

Why?

  1. I've always found difficulty to follow some training tutorials for yolo.
  2. This is supposed to help the beginners to train faster their models. You just download a dataset (or make your own), then run the trainer.
  3. It runs locally. That means you don't need internet or colab to train your models. Just a GPU.
  4. Because I'm using it to train some models.

Features

  • Easy Training: You can start training right after downloading a dataset in the following format:

    dataset (root)
|- test (folder)
|   |- labels (folder)
|   |- images (folder)
|- valid (folder)
|   |- labels (folder)
|   |- images (folder)
|- train (folder)
|   |- labels (folder)
|   |- images (folder)
|- data.yaml
    • TIP: You can download datasets from sites like Roboflow, or prepare your own dataset following the structure mentioned above.

  • Interactive Menu: After training the model, a user-friendly menu allows you to perform various tasks such as copying the result folder, copying only the model, testing the model, and generating another model.

  • Real-time Training: The program spawns a separate process to train the YOLOv5 model and displays training progress in real-time.

  • Model Testing: Test your trained model on custom images and visualize detection results.

Prerequisites

Before using the YOLOv5 Trainer, ensure you have the following dependencies installed:

  • Python (3.6 or later)
  • PyTorch (install through the PyTorch website to install CUDA support, if available)
  • OpenCv-Python (cv2)
  • PyYAML

Getting Started

  1. Clone this repository:

    git clone https://github.com/yourusername/yolov5-trainer.git
cd yolov5-trainer

  2. Instal requirements:

    pip install requirements.txt

  3. Run the trainer:

    python yolo_trainer.py

  4. Follow the trainer directions.

  5. When finished, you can copy the entire folder of the model, or just the model, and test it.

Helping

If you liked this repo, you can help me improve it. Just open a PR and I'll be looking into it.