Skip to content

README.md

Latest commit

 

History

History
63 lines (40 loc) · 1.98 KB

README.md

File metadata and controls

63 lines (40 loc) · 1.98 KB

​​ Create own Model and DataSet App - Great for learning (with GUI)

A desktop-based image classification app built using TensorFlow, OpenCV, and Tkinter. This user-friendly GUI tool allows you to create custom datasets, train models, and predict images—all without writing code.

Features

  • Create and manage image classes
  • Upload and auto-rename images into categorized folders
  • Train a CNN or pretrained model (e.g., MobileNetV2)
  • Predict new images with class probabilities
  • Correct misclassifications and add feedback to dataset
  • Fine-tune the model
  • Simple GUI via Tkinter for ease of use

Installation

1. Clone the Repository and Run Program

git clone https://github.com/jsoncodez/diyai.git
cd diyai


pip install -r requirements.txt


python diyai.py

💡 Usage Directions

<<<<<<< Updated upstream

  • Work in Progress...
image image image image image

  • Click "Train Model"

  • Click "Predict Image" and add image to run prediction.

  • Work in Progress...

💡 Usage Tips

  • Make sure each class has at least 10–20 images for decent results.
  • Use clear, centered images for better model accuracy.
  • You can correct a misclassified image after prediction, and the app will let you save it to the right class folder.
  • For better performance, retrain the model occasionally with newly added images.