TensorFlow outline

دوره: یادگیری عمیق با TensorFlow / فصل: TensorFlow - An introduction / درس 1

TensorFlow outline

توضیح مختصر

  • زمان مطالعه 0 دقیقه
  • سطح خیلی سخت

دانلود اپلیکیشن «زوم»

این درس را می‌توانید به بهترین شکل و با امکانات عالی در اپلیکیشن «زوم» بخوانید

دانلود اپلیکیشن «زوم»

فایل ویدیویی

متن انگلیسی درس

Pie again.

So far we only used num pi as a computational library but as you are probably guessing there are Python

libraries tailored for machine learning to our tensor flow and S.K.

learn in this section.

We will introduce tensor flow which will be the main tool we will use throughout this course.

Okay.

The mathematical concept of a tensor could be broadly explained in this way.

If a scalar has the lowest dimensionality and is followed by a vector and then by a matrix a tensor

would be the next object in line scalar as vectors and matrices are all tensor is of rank 0 1 and 2

respectively tenses are simply a generalization of the concepts we have seen so far.

We won’t get more into the mathematics.

I believe that’s all you need to know.

Once you see tensor flow you will quickly realize what I mean.

Okay.

Before introducing tensor flow it’ll be nice to take a minute and explain why it’s a good choice.

We will do so by comparing it to ask learn as they are two of the most popular libraries and most other

machine learning courses are based on SDK learn right.

You probably know it but Google is a leader in machine learning.

It is also one of the great innovators in the field because of the Google rain team.

As machine learning was developing Google needed better programming methods to suit their needs.

That is why they developed the tensor flow package for internal use.

At the end of 2015 Google released tensor flow to the public.

Currently it is probably the leading library for neural networks including deep neural networks convolution

neural networks and recurrent neural networks.

One of the biggest advantages of tensor flow is it uses not only the CPE of the computer but also as

GPO.

This is crucial for the speed of the algorithms as in this way tensor flow utilizes much more computing

power.

The best part is that this is done automatically.

Recently Google furthered this trend by introducing TPM use or tensor processing units which improves

performance even further tensor flow was one of the cutting edge technologies available right now and

is likely here to stay.

We set an alternative of tensor flow is SDK learn if you have done any machine learning but haven’t

used tensor flow.

You are probably familiar with the psychic learn or SDK learn library is very powerful and widely adopted.

However as Kane learned does not offer the same functionality as tensor flow regarding neural networks.

Having said that we can make the opposite point for other fields of machine learning in the presence

of problems such as K Means clustering and random forest esky alone could be a better fit even though

tensor flow started to make way.

That is especially true when it comes to pre processing as we will see later.

This part focuses on neural networks.

We aim to provide you with a deep understanding of neural networks instead of scratching the surface.

We decided to go deeper by introducing tensor flow in the lessons you will see next.

We will do just that.

We will start by introducing tensor flows underlying logic.

Then we will create the same minimal example but this time using the new syntax.

We will continue introducing machine learning theory and its implementation through tensor flow.

It is very important to know the theory is the same both for tensor flow as K learn or whatever package

you are using.

The only difference is the underlying code you’ll write.

That’s our outline.

Stay tuned.

مشارکت کنندگان در این صفحه

تا کنون فردی در بازسازی این صفحه مشارکت نداشته است.

🖊 شما نیز می‌توانید برای مشارکت در ترجمه‌ی این صفحه یا اصلاح متن انگلیسی، به این لینک مراجعه بفرمایید.