This repository contains a simple CNN training code using a small number of flower images. The code is provided without any personal information and is safe to share on GitHub.
- Create a folder named
FlowerImagesin the same directory as this repository. - Organize the folder structure as follows (create a folder for each class and put images inside):
FlowerImages/
daisy/
image1.jpg
image2.jpg
...
rose/
image1.jpg
image2.jpg
...
tulip/
image1.jpg
image2.jpg
...
- Supported image formats are
.jpgor.png. - Having at least 2 images per class is recommended for stable training.
pip install tensorflow matplotlibpython FlowerCNN.py- This will train the CNN using the training data and display graphs of training accuracy and loss.
- Currently, there is no validation data included.
- The code does not contain any personal information (e.g., user names or full paths).
- The image data is not included in this repository, so you must prepare your own
FlowerImagesfolder.