# Suggestions on how to better talk about PyTorch-Ignite by Sergey !

## What is PyTorch-Ignite ?

- Context (this and other bulleted line can be removed later)

Why using PyTorch without Ignite is suboptimal.
Ex: Pytorch is one of the leading ML libraries that makes ML engineer's life easier because: reasons. However (the scope of the project or whatever else) does not (what doesn't it do) that is where Ignite comes into play and makes the life of ML engineer or Data engineer or whoever else easier by: list of key selling points.

- Who are ignite's main users (and secondary users)
- What problems Ignite solves (can be development of the previous paragraph)
- what are ignite's values (what you think is important and what is less important)

> Ignite is a high-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently. For example, quickly train an image classifier using PyTorch and evaluate classifier's performance in terms of metrics (e.g. accuracy or precision).

- What are Ignite's key features

> Less code than pure PyTorch while ensuring maximum control and simplicity

> Library approach and no program's control inversion - Use ignite where and when you need

> Extensible API for metrics, experiment managers, and other components

- Example of a specific feature(s)
(or code to be used in more introductory presentations. So that the user will see a concrete example of how life with Ignite is better)

## Ignite's place in the ecosystem

(the subtitles should be changed later. At the moment they are just guidelines, not for publication)
- What are Ignite's competitors and how it supasses them
- What are the applications where Ignite is the best
- What are Ignite's features that others haven't implemented yet (not copied yet)

## Roadmap

**(continue to be the best at X, Y, improve Z, explore the possibility of M, N)**

> [name=me] We can talk about this: https://github.com/pytorch/ignite/wiki/Roadmap

> [name=Sergey Epifanov] right, but more from the functional perspective. Not how you manage the organizations, but what features the users should expect.

> [name=me] this is something to be careful about
>
> [name=Sergey Epifanov] true. that's why you don't need to be specific, no dates, no promises.

- where Ignite is going
- what areas the team plans to strengthen
- what new usages of Ignite should be explored