From the docker folder, run this command: docker-compose up -d --build --remove-orphans

Then navigate to localhost:20000 to access the jupyter notebook.

Each folder in the src directory is for a different ML task. The purpose of this repository is for me to experiment with different ML tasks using one docker image having all dependencies installed.

https://www.kaggle.com/c/titanic/data OR kaggle competitions download -c titanic

https://www.kaggle.com/ngbolin/mnist-dataset-digit-recognizer/data

https://www.kaggle.com/shivamb/how-autoencoders-work-intro-and-usecases/data

Downloads the data when you run the script

Clone: [email protected]:learning1234embed/NeuralWeightVirtualization.git

Run: ./download_dataset.sh && python weight_virtualization.py -mode=a -network_path=mnist

Copy the contents of the mnist folder except the .py(c) files