First and foremost, join us at GANs Workshop Slack Channel so that we can keep in sync. Most probably, this channel will be the starting point for a local community of developers interested in hands-on AI & ML stuff.
For this workshop we will use the MNIST and Celeba datasets. You can download them from this google drive folder. The folder contains the two img_align_celeba
and mnist
ziped folders. Please create a data
folder in the project repository's root directory, copy the two dataset folders there and extract them.
The final paths relative to the project root should be: ./data/img_align_celeba
and ./data/mnist
.
We will be using the Anaconda platform for this workshop. Anaconda is a distribution of packages built for data science that comes with conda, a CLI packages and environments manager.
- Install Conda CLI: we are mainly interested in conda, therefore you have two options:
- Install Miniconda - a mini version of Anaconda that includes only conda and its dependencies
- Install the entire Anaconda ecosystem which contains conda plus over 720 open source packages
- Download the Python 3.6 installer for your OS and follow the instructions from here.
- Update all the packages in the default root environment:
conda upgrade --all
If you get "conda command not found" add export PATH="/Users/username/anaconda/bin:$PATH"
to your bash config file.
Most used commands:
conda install package_name
conda update package_name
conda remove package_name
conda search search_term
conda list
- list all installed packages
Most used commands:
conda create -n env_name list_of_packages
conda env export > environment.yaml
- save packages to .yaml fileconda env create -f environment.yaml
- create env from yaml filesource activate my_env
(Mac/Linux) ORactivate my_env
(Windows) - activate environmentsource deactivate
(Mac/Linux) ORdeactivate
(Windows) - deactivate environmentconda env list
- list all environmentsconda env remove -n env_name
For more info check this and this
For this workshop create a new environment using the provided gans_workshop.yaml by running:
conda env create -f .environments/gans_environment_OS.yml
Activate the new created environment using:
source activate gans_workshop
(Mac/Linux)activate gans_workshop
(Windows)
Then run the following command to start a test jupyter notebook:
jupyter notebook GANsWorkshop.ipynb