MortenTabaka / Deep-learning-exercises-with-TensorFlow2

The repository contains three deep learning models created for Kaggle and PASCAL datasets.

Machine learning excersises

1. Download datasets

All datasets must be downloaded to ./data/

Run mkdir data in working directory or in Python:

if not os.path.exists('./data'):
    os.mkdir('./data')

• Ex.1 notebook for Multi-class classification in TensorFlow2. Polish craft beer labels dataset

• Ex.2 notebook for Flowers recognition with augmentation techniques. Flowers recognition dataset

• Ex.3 notebook for Multilabel classification PASCAL VOC2007. Pascal VOC2007. To download and unzip dataset do_download=True in the Tabaka_Marcin_Transfer_learning.ipynb.

• Ex.4 notebook for Augmentation techniques in TensorFlow2 on image (included in repository)

2. Create Anaconda Environment

• To install Anaconda follow Anaconda documentation
• Open terminal in project directory and run conda env create --name CHOSEN_ENV_NAME --file tf2.yml or conda create --name CHOSEN_ENV_NAME --file pkgs.txt
• Run environment by conda activate CHOSEN_ENV_NAME
• Run Jupyter by jupyter-notebook
• Open and run notebooks