Past Meetup

3 hour workshop: Deep Learning + DevOps with TensorFlow and Docker

This Meetup is past

462 people went

Location visible to members

Details

Costs : Free of charge

Github Code:

https://github.com/QuantScientist/deep-ml-meetups/blob/master/tensorflow-dl/README.md

Summary:

In this completely demo-based session we offer 3 hours of free instruction on using Google’s TensorFlow (https://www.tensorflow.org/) in a cloud based env. Participants will setup a google cloud image, install TensorFlow (TF) on a Docker image and code basic machine learning algorithms.

Are there any prerequisites?

Note: All demos are 100% open source using Jupyter Notebook, Docker, and google data lab. Therefore, participants must bring their on laptop and have a full docker installation. No theory will be presented.

Previous experience programming in Python or in other languages is advised to make best use of the workshop. Additionally, some familiarity with machine learning is necessary.

Tensorflow supports both Python 2.7 and Python 3.3+. Note that for Windows, TensorFlow supports only 64-bit Python 3.5. For this session, I will use Python 2.7. But you’re welcome to use either Python 2 or Python 3.

Google has a pretty detailed instruction on how to download and setup Tensorflow. You can follow it here: https://www.tensorflow.org/get_started/os_setup

Unless your computer has GPU, you should install Tensorflow without GPU support.

Tentative Schedule: 18:00 to 21:00

• Mac OSX TensorFlow Docker Sandbox (Shlomo Kashani)

• TensorFlow and BigQuery . An end-to-end example of using Tensorflow. (Aviv Rotman and Jenia Gorokhovsky )

In this session we'll get our hands dirty and build an end-to-end Deep Learning pipeline. We will:

- Query and process data using Google BigQuery

- Pipeline data directly to TensorFlow and build a simple model

- Monitor training using Tensor Board

- Export a model for serving and serve online requests

By the end of the session we should have a modest real-world model ready for production.

Aviv Rotman

Aviv is a Data Scientist at Taboola. He holds an MSc. in Biology from the Weizmann Institute of Science. Since joining Taboola Aviv spends most of his day fiddling with models in TensorFlow and trying figure out how to deliver meaningful metrics and insights using BigQuery.

Jenia Gorokhovsky

Jenia is an Algorithms Developer at Taboola with over 8 years of experience working with Machine Learning and data pipelines. At Taboola Jenia spends much of his time moving the main Machine Learning pipeline to TensorFlow.

Prerequisites:

Setup required prior to the meeting:

1) Install Docker Engine (https://docs.docker.com/engine/installation/) On Your Environment (Linux, MacOS, Windows):

2) On Your Environment, Download Docker Image and Start Docker Container in the Background:

https://github.com/QuantScientist/deep-ml-meetups/tree/master/nice-docker

``` docker pull quantscientist/deep-ml-meetups ```

``` docker run -it -p 8888:8888 -v /my-host-data-science-folder/:/root/sharedfolder quantscientist/deep-ml-meetups bash ```

3) On Your Environment, Shell into the Docker Container:

``` chmod +x run_jupyter.sh

./run_jupyter.sh

```

Github code:

https://github.com/QuantScientist/deep-ml-meetups

https://github.com/QuantScientist/deep-ml-meetups/blob/master/tensorflow-dl/ (https://github.com/QuantScientist/deep-ml-meetups/blob/master/tensorflow-dl/README.md)

References:

https://www.slideshare.net/matthiasfeys/introduction-to-tensorflow-66591270

https://jrmeyer.github.io/tutorial/2016/02/01/TensorFlow-Tutorial.html (https://www.slideshare.net/matthiasfeys/introduction-to-tensorflow-66591270)

https://www.tensorflow.org/get_started/get_started (https://www.slideshare.net/matthiasfeys/introduction-to-tensorflow-66591270)

https://github.com/aymericdamien/TensorFlow-Examples (https://www.slideshare.net/matthiasfeys/introduction-to-tensorflow-66591270)