Fast.ai 2019 Deep Learning - Lesson 5

Lesson no. 5 for our study group.

Please remember to complete online video, code and homework for this lesson before the meeting!

More info and schedule on wiki: https://wiki.hs3.pl/wiki/datascience#kurs_deep_learning_-_fastai

---

In this study group, we will be working through the fast.ai deep learning course "Practical Deep Learning For Coders v3, Part 1": https://course.fast.ai/ .

This meetup is free and open to all.

Over the next 14 weeks we will be following the 2019 version of the fast.ai course, meeting every 2 weeks.

In each session we will use that week's fast.ai lectures and course materials as a basis for discussion and learning. Everyone is invited to contribute their insights and questions.

Prior to each session watch the lecture for that week and work on course assignments. In this session we cover image classification.

The fast.ai course is based around Python 3.6, so basic familiarity with python is a plus. For the deep learning component, fast.ai supplies its own package (fastai) which is built on top of PyTorch, a python package for tensor computation and deep learning.

About the course:

7 lessons in Part 1 (about 20 hours of video)

1 - Image classification

2 - Production; SGD from scratch

3 - Multi-label; Segmentation

4 - NLP; Tabular data; Recsys

5 - Backprop; Neural net from scratch

6 - CNN deep dive; Ethics

7 - Resnet; U-net; GANs

Expect to spend around 5 hours per lesson of your own time (i.e. to watch the video and run the homework programs).

Prerequisites: Basic coding, math concepts are introduced as needed.

Authors suggest to have at least 1 year of prior coding experience in any language: https://course.fast.ai/#getting-started

Where to run lessons:

- Personal computer with Nvidia GPU. CPU option possible but not optimal

- GPU enabled Cloud Virtual Machine

https://course.fast.ai/index.html#using-a-gpu

Feel free to contact us with any questions.

Please note that we have no official connection with fast.ai.

See you there,

Hackerspace Trójmiasto Data Science Group

Amadeusz Lisiecki

Jakub Kruszyński

Michał Wojczulis

Robert Różański