Overview

This repository is meant to facilitate an interactive quick introduction to deep learning. If you have stumbled upon this repo on your own, there are other better resources for you on the internet.

@Datanauts add your questions, thoughts, concerns, resources in this community wiki If you want to contribute, please send me a pull request!

Setup

Tutorial Setup

To build your own Docker image, see Build Setup

• Install Docker https://docs.docker.com/engine/installation/

• If you are using NVIDIA CUDA-enabled GPU, install nvidia-docker

• Pull Docker image:

  • CPU (Windows):

    sudo docker pull alexisylchan/introdlcpu-windows

  • CPU (Other operating systems):

    sudo docker pull alexisylchan/introdlcpu

  • GPU (CUDA 7.5):

    sudo docker pull alexisylchan/introdlcuda7.5

  • GPU (CUDA 8):

    sudo docker pull alexisylchan/introdlcuda8

• Install Git-LFS for Linux, Windows, or Mac

• Clone this repository

  git clone --recursive git@github.com:alexisylchan/introdl.git

• Run bash in your Docker container

  • If you're using CPU:

    chmod 777 start.sh

    • Windows:

      ```docker run -it -p 8888:8888 -p 6006:6006 <docker_image> bash```
      

    • Non-windows:

      ```./start.sh <absolute_path_to_introdl_folder> <docker_image>```  
      

  • If you're using GPU:

    -it -p 8888:8888 -p 6006:6006 -v <absolute_path_to_introdl_folder>:/root/introdl <docker_image> bash```
    
    

• Start Jupyter within the Docker container

  jupyter notebook

Build Setup

• Install Docker https://docs.docker.com/engine/installation/

• If you are using NVIDIA CUDA-enabled GPU, install nvidia-docker

• Install Git-LFS for Linux, Windows, or Mac

• Clone this repository

  git clone --recursive git@github.com:alexisylchan/introdl.git

• In the introdl directory, build your Docker image. This step may take 1-2 hours

  • CPU:

    • Windows:

      sudo docker build -t <your_user_name>/introdl -f Dockerfile.cpu-windows .

    • Other operating systems:

      sudo docker build -t <your_user_name>/introdl -f Dockerfile.cpu .

  • Machine with NVIDIA CUDA-enabled GPU:

    ```sudo docker build -t <your_user_name>/introdl -f Dockerfile.gpu .```    
    

• Install Google Chrome (I had issues running Jupyter notebook on Firefox within the Docker container)

  • On Ubuntu, you may encounter installation issues which require enabling the Universe Repository

• Run bash in your Docker container

  • If you're using CPU:

    chmod 777 start.sh

    • Windows:

      ```docker run -it -p 8888:8888 -p 6006:6006 <docker_image> bash```
      

    • Non-windows:

      ```./start.sh <absolute_path_to_introdl_folder> <docker_image>```  
      

  • If you're using GPU:

    -it -p 8888:8888 -p 6006:6006 -v <absolute_path_to_introdl_folder>:/root/introdl <docker_image> bash```
    
    

• Start Jupyter within the Docker container

  jupyter notebook

Attribution

Thank you to the following resources:

• saiprashanths/dl-docker for the deep learning docker file
• Jason Brownlee's excellent time series prediction with deep learning in Keras
• Yhat's blog post on moving averages
• Logan Engstrom's Fast Style Transfer
• NASA APOD

(Please let me know if you would like to be added to this list!)