Titanic on GCP ML-Engine
Code me like one of your french girls...
This repo is a demo for using GCP ML-Engine to train a scikit learn model on GCP ML-Engine
Setup
- Install Python in a virtual environment and then install and configured the GCP Cloud SDK
- Install all python requirements
pip install -r requirements.txt
- Create a storage bucket using
create_bucket.sh
- Download train.csv and test.csv from Kaggle's Titanic competition and place it in /data in this repo
- Run data_prep.py
data_prep.py
This script downloads the 'train.csv' file hosted in the GCP project bucket and then creates a training and validation dataset using that data. Finally the training and validation dataset are both uploaded back to the GCP project bucket