Name: TensorFlowのハイレベルAPIを5分で使いこなそう｜Google I/O 2018 (Get started with TensorFlow's High-Level APIs in 5 mins | Google I/O 2018)
Uploaded: 2021-01-14T10:33:28.000Z
Duration: 5 min 15 s
Description: VoiceTubeの動画で発音を聞きながら英語表現を覚えよう！学べる英語：

to 不定詞

we're here today to speak with you about tensor flows High level AP ice So one area and particularly passionate about is making machine learning as successful.

基礎数量詞

It's possible to as many people as possible, and the Tensorflow team has been investing very heavily in the same thing.

過去形

So we spent a lot of energy making tensorflow easier to use.

So there's three things that are concrete that I'd like to walk you through.

And the very first is even if your brand new to tensorflow your brand new machine learning even if you're new to Python.

One area that seems silly but is nontrivial for a lot of people is actually just installing tensorflow indifferent dependencies.

And I know for Python developers is just pip install tensorflow.

But that could be hard for people that are brand new.

So I'm going to show you something called Collab, and I'll walk you through KOA Lab.

It's basically a Jupiter notebook server running the cloud.

It's free of charge as Tensorflow preinstalled comes with free GPU.

The next thing tensorflow as many different AP eyes but my personal favorite What I'd strongly strongly recommend to you is something called Care us and care us care as a p.

I is completely implemented inside of Tensorflow.

I can't tell you how much fun I've had using it.

So I'll walk you through writing Hello world and care us the same A p I is also useful for tensorflow Js and then I'm to point you to some educational resource is to learn more eso briefly.

This is what I would do if you want to try tensorflow and care Ross and TF data and eager execution in the fastest possible way and I should tell you off the bat.

So all these ap eyes, they're fully implemented and they're working well.

We are just now starting to write all the samples and docks around them.

So I have a feeling the samples I was able to cook up for this talk.

But stay tuned and check back in the next few months is we flush this out.

But let me just show you how to dive right in.

So if you go to this website, it will bring you to this getup site.

And if you scroll down to the read me, you'll see a sequence of a few notebooks, and I just want to show you how easy it is to get started if you just click on one.

What happens is they open up immediately in Collab.

It's running entirely in the cloud you can hit Connect to connect to a colonel.

And now I can start running these cells, and I'll walk you through this in more detail in a few minutes.

But if you go through the first notebook, this is going to show you how to write your first neural network using care us.

Um, there's a little bit of pre processing code, but the notebook is very short.

The next notebook will show you how to do the same thing.

Using care, awesome combination with TF data and eager execution.

And then we go into a little bit more depth.

So it's literally that easy to get started.

So there's broadly five steps to write Hello world in Tensorflow using care us.

The good news is steps 34 and five are literally one line of code, and you can see that on the second line we're importing M n'est, and this is easy because the data set is already we have a loader for it that's baked into tensorflow.

Here's the format of the data set, so as imported, it's divided already for us into train and test train is about 60,000.

I have a diagram of the format of the images.

If you look at the notebooks in that workshop directory, the best thing you can do when you imported data set is to spend a lot of time asking really basic questions so literally when you import the data printed out and there's a couple points that I want to make one.

This is the complete code to define the network, so it's code concise too broadly.

The more layers you add to your network, and the more neurons air units per layer, the more capacity your network has meaning.

The more types of patterns that can recognize the problem is the more things your network can recognize, the more likely it is to memorize the training data.

Building your model is where there are many, many machine learning concepts that you have to spend a lot of time learning the next three steps.

They're literally concepts that are basically you've involved with running an experiment.

So it's one line fit is synonymous with train, and we're training at using the training images in the training labels.

Training a network is a little bit like tuning a guitar.

So think if you have a guitar and you want, it starts untutored and you want to tune the string instead of particular note.

So you start tuning it and like every time you twist the wheel on the guitar, you can think about is an F block.

After that, you can evaluate it and evaluate just means, given some new data classified with my network.

And take a look at the accuracy and other metrics that's also just one line of code.

 Everyone really appreciate your time and hope this stuff is useful to you. 