fastai-studygroup-multiinput

fastai study group: Multi Input Practical deep learning have data from various sources, some from structured data, attributes, time series data and images. Multi-Input (Mixed input) is to combine all of these types into one solution. A common case will have images, time series, and text meta-data, could be process that takes images and report attributes about the situation the images taken. Share market analysis as well, will have time series data and attributes like sentiment or market conditions.

Fast ai introduced mixed-input in lesson 3/4 in the Rossmann store price prediction. The data is taken from the lesson, after processing the text files. The problem presented is mixed input of tabular data with continuous variables and text meta-data. The study group forum in fast.ai: http://forums.fast.ai/t/study-group-in-perth-australia/22556 The project is designed to run in colab, so please get your account ready. Download the following files:

1. https://drive.google.com/open?id=19oXlEAkcCrylZc9QvoioxaBhl-1Jtnha
2. https://drive.google.com/open?id=1DQVXvAN_ZTM4Esy4gBKn817IiV5F0Jez
3. https://drive.google.com/open?id=1rtyf5sZzSmmEuhwBCr8ta1WCfcYgg4Vf

This project will run in google colab: https://colab.research.google.com If you have a google account, you can run the note book