Track 1 (18.11.2018)
Building Interactive Dashboards in Python - First steps with Dash - Mikolaj Olszewski
This tutorial will be highly interactive. Participants are advised to bring their own laptops with the following software installed:
- Python 3 with packages:
- pandas
- numpy
- dash
- dash-html-components
- dash-core-components
- dash-table
- plotly
- Text editor of you choice (for example Atom)
If you had those packages installed, please make sure they're up-to-date. Please also make sure that you're able to run Python scripts from the command line.
Recognize drawings in the browser with Tensorflow.js - Karol Majek, Monika Koprowska
Prerequisites for this tutorial:
- Keras, Tensorflow - python, basic knowledge
- Google Chrome or Firefox browser
- up to date
- WebGL support:
https://get.webgl.org/ - Have non zero predictions in this demo:
https://storage.googleapis.com/tfjs-examples/mobilenet/dist/index.html- GitHub account
- We will use GitHub Pages (
github.io)
- Google Colab or Jupyter notebook familiarity
- We will train model in Colab
- Google account is not required
- Git client is not required
Track 2 (18.11.2018)
Structuring machine learning models by using pipelines - Paweł Jankiewicz
This tutorial assumes scikit-learn familarity and writing simple classes in Python.
The materials will be presented using Jupyter Notebook so please prepare an environment with standard Python data science stack:
- scikit-learn
- pandas
Serverless Approach to Working with Data - Jakub Nowacki
To run this repo you'll need:
- Install Node.js https://nodejs.org/
- Install Serverless https://serverless.com/framework/docs/getting-started/
- Install Python distribution, e.g. Anaconda https://www.anaconda.com/download/
- Install your favorite Python ID, e.g. VSCode https://code.visualstudio.com/; support for JavaScript, NodeJS, YAML etc. is a plus.
- Install Git https://git-scm.com/
- Register on AWS https://aws.amazon.com/ and install AWS CLI; you should not surpass the free tier.
- For Windows 10 users it is useful to have Ubuntu for Windows https://tutorials.ubuntu.com/tutorial/tutorial-ubuntu-on-windows#0
- For deployment package builder it is good to have Docker installed (https://www.docker.com/get-started) along with Serverless Python Requirements Plugin (https://www.npmjs.com/package/serverless-python-requirements)
Resolving Docker Toolbox Daemon Is Not Running Error (Windows)
The way to fix Docker Toolbox daemon error is to set a number of environment variables, as follows:
SET DOCKER_TLS_VERIFY=1
SET DOCKER_HOST=tcp://192.168.99.100:2376
SET DOCKER_CERT_PATH=%USERPROFILE%\.docker\machine\machines\default
SET DOCKER_MACHINE_NAME=default
SET COMPOSE_CONVERT_WINDOWS_PATHS=true
%USERPROFILE% should point to your home directory (you can check it using echo %USERPROFILE%); if it is not set correctly, change manually to your home in DOCKER_CERT_PATH given above.
The description is taken from https://www.mydatahack.com/resolving-docker-deamon-is-not-running-error-from-command-prompt/
Track 3 (21.11.2018)
Introduction to Recommendation Systems - Piotr Bigaj, Jakub Gasiewski, Przemek Kepczynski
Requirments:
- Python 3.6+
- Anaconda (numpy, pandas, jupyter notebook)
- Tensorflow (if there will be enough time to use WALS)