Download Subtitles and Closed Captions (CC) from YouTube

Enter the URL of the YouTube video to download subtitles in many different formats and languages.

TensorFlow Swift with English - CC subtitles   Complain


MAGNUS HYTTSTEN: Hi, I'm Magnus Hyttsten for the Developers

Show, and we are here at the TensorFlow Dev Summit.

And I'm standing here with Chris Lattner.

Hello, Chris.


MAGNUS HYTTSTEN: So Chris, you are actually

one of the original founders of the Swift language.

Is that correct?

CHRIS LATTNER: That's true.

It started as a nights and weekends project,

and it grew to something a little bit bigger than that.

MAGNUS HYTTSTEN: That's absolutely amazing.

And you just had a talk on TensorFlow and Swift

that may be surprising to some people.

How does these two things-- how did they get together?

CHRIS LATTNER: Well, it turns out

that Swift has some really interesting capabilities

that most other languages don't have.

And so one of the neat things about Swift

is it was designed from the very beginning

with the idea of spanning all the way from systems

programming all the way up to scripting languages.

And it turns out that TensorFlow is really

a sweet spot for that, because you

want to be able to program TensorFlow like a scripting

language and be able to use high-level abstractions,

but you also care about low-level performance as well.

And so Swift is really well-positioned to fit that.

Beyond that, Swift is a great language

because it supports the kind of compiler analysis

and all the deep internal stuff that

makes it possible to do what we're doing

with Swift for TensorFlow.

MAGNUS HYTTSTEN: And it's very modern in syntax, compact,

and it has the construct that you would expect?

CHRIS LATTNER: Absolutely.

So Swift was starting in 2010, so it's not

that old by language standards.

And so one of the great things about that is it

gets to build on all the best things that have been learned,

that are modern techniques in programming languages.


So I think most people-- or you developers

out there-- that you associate Swift with iOS development.

But this has nothing to do with iOS development, correct?

CHRIS LATTNER: Absolutely.

Swift is an open source language,

and people use it for server development

and a lot of other things that are beyond iOS as well.

And so this is just a great way to express tensor computations

and be able to build your machine learning algorithms

in a natural way where the compiler can

help you write your code.

It can detect errors for you.

It can make your life better so that you

can focus on getting the job done instead

of trying to fight through the framework

and trying to deal with typical problems.


So I guess the question on everyone's mind

is how can I get involved in this?

How can I use Swift with TensorFlow?

CHRIS LATTNER: Oh, that's really easy.

So it's still a very early stage project,

but we're opening our open source project

in the next month.

So in the month of April, you can sign up.

At that time, we're publishing not just the code,

but also white papers.

And we're also allowing open development,

so you too can contribute to the project.

And so if you're interested in that,

you can go to ommunities/swift.


So it's going to be really interesting to see

what you developers create with Swift and TensorFlow.

I'm Magnus Hyttsten.

This is Chris Lattner.

We're at the TensorFlow Dev Summit for Developers Show.


Download Subtitles Download Video Download Audio

↑ Return to Top ↑ - contextual dictionary