PyData DC 2018

Skorch - A Union of Scikit-learn and PyTorch

Presentation

The slides can be downloaded at: download link.

To clone this repo run:

git clone --depth 1 https://github.com/thomasjpfan/pydata2018_dc_skorch

Setup

To run the notebook locally, please following the following setup procedure:

1. Install dependencies: conda env create -n pydata_dc_2018 -f environment.yml
2. Activate env: conda activate pydata_dc_2018

Part 3

1. Follow Kaggle's installation and configuration documentation to install and configure the kaggle cli
2. Go to Kaggle's 2018 Data Science Bowl Competition, click on "Late Submission" and accept the terms and conditions to get access to the data.
3. Run ./dl_extract_prepare.sh to download, extract and prepare the data.

Run Jupyter Lab

1. Activate env: conda activate pydata_dc_2018
2. Launch jupyter lab: jupyter lab