Brief presentation on fast.ai -- what would you include?

memetzgz · April 3, 2018, 2:00pm

Hello all you smart people

I’ve signed myself up to do a short (7 min. max) talk on the fast.ai library and course at a local PyData meet up next week.

My goal is to give an overview of the benefits of the library and what’s covered in the two parts of the course, and then provide links so people who are interested can investigate more (and then hopefully take the course and use the library, of course!) My audience is likely more familiar with machine learning than deep learning.

I’ve got some ideas of my own, but since it’s become clear from the class discussion and the forums that you all are way smarter than I am, I thought I would ask you all what you would include in such a presentation.

Remember, I’ve only got 7 minutes . . .

All suggestions welcome!

gerardo · April 4, 2018, 3:17am

Cats & Dogs - Transfer Learning
https://colab.research.google.com/drive/1YZlftYhbHEbmb9AesIBbiXE8noTU2BPI

anshbansal · April 4, 2018, 5:01am

A link to the abbreviations list at the start or end. A lot of people would be confused with the short hand used in the library. The abbreviations list would help.

github.com

fastai/fastai/blob/master/docs/abbr.md

## Fastai Abbreviation Guide
As mentioned in the [fastai style](https://github.com/fastai/fastai/blob/master/docs/style.md), 
we name symbols following the *Huffman Coding* principle, which basically means 

> Commonly used and generic concepts should be named shorter. You shouldn't waste short sequences on less common concepts. 

Fastai also follows the life-cycle naming principle:
> The shorter life a symbol, the shorter name it should have.

which means:
- **Aggresive Abbreviations** are used in *list comprehensions*, *lambda* functions, *local helper* functions.
- **Aggresive Abbreviations** are sometimes used for *local temporary variables* inside a function.
- **Common Abbreviations** are used most elsewhere, especially for *function arguments*, *function names*, and *variables*
- **Light or No Abbreviations** are used for *module names*, *class names* or *constructor methods*, since they basically live forever. 
However, when a class or module is very popular, we could consider using abbreviations to shorten its name.

This document lists abbreviations of common concepts that are consistently used across the whole fastai project. For naming of 
domain-specific concepts, you should check their corresponding module documentations. Concepts are grouped and listed by semantic order. 
Note that there are always exceptions, especially when we try to comply with the naming convention in a library.

This file has been truncated. show original

and link to some interesting blogs that people might have written to use it. Code examples are helpful when I want to start using a new library.

DavideBoschetto · April 4, 2018, 8:32am

Wow, congratulations!
I’d focus on the ease of use of the library! To do this, the “7 lines of code” mantra of Part 1 would be the best approach in my opinion.
In 7 minutes, and for an audience not expert in deep learning, you could be able to convey the usability-message (“state of the art results in one page of code!”).

Edit: dogs and cats might be the example that most people will point you to, but if your field is EHR I’d focus on note classification/language models! Trying to explain the Rossmann competition in 7 minutes might be fun…

PrinceP · July 9, 2018, 1:27am