Flower Classification with PyTorch

Project Overview

This project is part of the our University's course "Introduction to Deep Learning". The goal of this project is to classify flowers using CNN and 3 different pretrained models (ResNet, DenseNet, EfficientNet). The dataset used is the Flower Classification | 10 Classes | dataset from Kaggle.

• You can find more information about the project in the About Project page.

Project Usage (only streamlit page)

1. Clone the repository

git clone <repo-url>

2. Install the requirements

pip install -r requirements.txt

3. Run the streamlit page

streamlit run Main_Page_🏡.py

Project Usage (training)

1. Clone the repository

git clone <repo-url>

2. Get the kaggle key

• Go to your kaggle account and download the kaggle.json file. Then, move it to the src folder. You can also follow the instructions here

3. Go to the src folder

cd src

4. Create the conda environment

• If you don't have conda installed, you can follow the instructions here. I recommend using miniconda.

conda env create -f environment.yml

5. Activate the conda environment

conda activate torch

6. Run the training script with the desired options

• You can find different training options in the src/utils/options.py file
• You don't have to specify all the options and also don't have to download the dataset. If you place kaggle.json in the src folder, the script will download the dataset automatically.

python main.py --option1 value1 --option2 value2 ...

Contributors

• Eren Akgül - GitHub | LinkedIn | email
• Yasin Tarakçı - GitHub | LinkedIn | email