Tensors Explained - Data Structures of Deep Learning

[@pedro1492]

finally an explanation slow enough for my smooth brain to keep up

[@doomleika]

This is great, many of the docs are very bad at clearing up the terms and this clear things up really good

[@benphua]

Great intro to the subject, I was confusing it with the physics definition of a tensor, much appreciated.

[@manalshamseldien4389]

I really wanna thank all of the team , The basics now of tensors are firmly in my mind

[@ahmedaj2000]

i love how you compared the cs and math sides, great explanation! thanks!

[@plus1knowledge246]

The best description I've heard of a tensor, thanks!

[@深夜酒吧]

I am enjoying the music at end of this video. Never had this experience from other KZbin channels.

[@whoisbhauji]

Thanks for the simple explanation - couldn't understand what some 'sponsored' videos were trying to explain in long videos

[@vxsniffer]

this series is the best nn introduction I've seen so far... thank you and keep going!

[@deeplizard]

Thank you! Appreciate that!

[@ravihammond]

I've always wondered what tensor means... I am now enlightened. You are the man.

[@deeplizard]

Moving along!

[@tymo3391]

High quality production! Thank you

[@deeplizard]

Check out the corresponding blog and other resources for this video at: deeplizard.com/learn/video/Csa5R12jYRg Jeremy's Ted talk: kzbin.info/www/bejne/qmXOqoWvgN6CptE

[@余明新]

you use the wrong noun "dimension''.you should use the word "rank".

[@misandaibo6186]

{ "question": "Rename variable y: y = [[4, 2], [3,1], [0, 8], [6, 1]] ", "choices": [ "y is a 2D tensor", "y is a 1D array", "y is a 2D vector", "y is a 4D tensor" ], "answer": "y is a 2D tensor", "creator": "Misan", "creationDate": "2020-08-16T20:42:11.824Z" }

[@deeplizard]

Hey Misan - Thank you for making this question. I added it to the quiz 😊 - Chris

[@Brahma2012]

Excellent intro to the concept of a Tensor..Thanks

[@richarda1630]

Great explanation for us noobs! The background music accents are a nice touch !

[@Diamond_Hanz]

Thank you. I like your high level explanation of this! All the other people jus toss words around.

[@MMphego]

I have been enlightened... Ya'll are doing the lords work. Thank you

[@philipweslow565]

{ "question": "How many indices are necessary to access a specific element of the tensor \"z\"? z = [7, 8, 9, 10]", "choices": [ "4", "3", "2", "1" ], "answer": "1", "creator": "PW", "creationDate": "2019-07-05T21:06:21.999Z" }

[@deeplizard]

Thanks Philip! This one is added. 🧠 🚀

[@sofeint6050]

amazing video, so simple, exaclty what i needed

[@sibabratabiswal1064]

we will talk about the content of the video later...firstly great bg music...great story telling and great video ...content , great explantation

[@sandromartins3579]

Thanks for making it simple

[@aravindk4967]

At 4:31, does that tensor not have only 2 components? Aren't components just the coordinates or indices? How are tensor dimensions different from dimensions in vector space?

[@deeplizard]

The components or component values are the scalar values (numbers) inside the tensor or vector. The tensor at 4:31 has two axes (rank 2) and has 9 component values. The number of dimensions a tensor has depends on the number of axes inside the tensor. The number of dimensions a vector has depends on the number of components inside the vector.

[@aravindk4967]

@@deeplizard Right, I think I understand it now. Thanks a lot!

[@vinodhkumarRajasekaran]

Wonderful Explanation !

[@reemal-fayez9900]

Thank you!

[@benevolentbear8213]

you guys deserve a lot more recognition. Thanks for these amazing videos.

[@sulemanrasheed1634]

Great work guys! Keep it up

[@lonewolf2547]

You guyz are brilliant

[@magelauditore333]

Awesome explanation. I came from keras tutorial by Mandy. And now pytorch really loving it. Keep it up keep going. And the music is kind of creepy but i kind of like it

[@philipweslow565]

{ "question": "How many dimensions does the tensor \"z\" have? z = [7, 8, 9, 10]", "choices": [ "0 dimensions.", "1 dimension.", "2 dimensions.", "4 dimensions." ], "answer": "1 dimension.", "creator": "PW", "creationDate": "2019-07-05T21:09:41.517Z" }

[@deeplizard]

Thanks Philip! This one is added. 🧠 🚀

[@philipweslow565]

@@deeplizard My pleasure!

[@twitteranalyticsbyad3969]

Lets get started and there comes the 6 seconds dragon. nice... 00:00:22

[@StopWhining491]

Never heard the term "tensor." Now I sort of know why. Also enjoyed the video. "Would you like to play a game?"

[@shanakaj007]

best explanation

[@philipweslow565]

{ "question": "Which of the following would qualify as a tensor?", "choices": [ "A 2D-array.", "A matrix.", "A vector.", "All of the above." ], "answer": "All of the above.", "creator": "PW", "creationDate": "2019-07-05T20:59:20.285Z" }

[@deeplizard]

Thanks Philip! This one is added. 🧠 🚀

[@PrashantSharma-ql4yb]

Great video and Excellent Website, Subscribed! Just a suggestion, you might want to get an SSL Cert for your site, all browsers these days notify users of connection reliability and your website could be presumed as unsafe by some novice visitors.

[@deeplizard]

Thank you. I appreciate that. Also, appreciate you taking the time to mention the enhancement to the site. I've been planning to get SSL added. AWS didn't provide a straight forward way to implement SSL when I set the site up. It's about time a circle back around and take care of it.

[@deeplizard]

Got the site switched over to https. Thanks again for your concern regarding this.

[@eug_gino]

thank you sooooomuch

[@deeplizard]

You are welcome!

[@imacprousersam7306]

Easier than I thought 💭

[@aaryanvaidya3501]

godsend video

[@AliAwadh980]

Hi, thanks very much for your wonderful videos. I was wondering if you can share how to convert text to semi-natural speaking voice, or if you can share few links that I can read.

[@deeplizard]

Hey Ali - You are welcome! The big cloud providers offer APIs for this, and they have free tiers. Google: cloud.google.com/text-to-speech/ IBM: www.ibm.com/watson/services/text-to-speech/ Amazon: aws.amazon.com/polly/ Microsoft: azure.microsoft.com/en-us/services/cognitive-services/text-to-speech/ Here is one that looks interesting on GitHub: github.com/mozilla/TTS FYI: We've never used any of these. If you try any of them or find another solution. We'd love to hear about your experience. Good luck!

[@AliAwadh980]

Thats awesome! Thank you very much. I was looking much for open source, so I'll focus on TTS from mozilla for now. Thank you again Deeplizard!

[@louerleseigneur4532]

merci merci

[@ninjaturtle205]

is this a part of a series? Where is the rest?

[@deeplizard]

deeplizard.com/learn/playlist/PLZbbT5o_s2xrfNyHZsM6ufI0iZENK9xgG

[@iamzainshahbaz]

THANK YOU SO MUCH

[@deeplizard]

No problem 😊

[@reyboyvideogames]

This is so well explained put this on udemy plx

[@kevinayers7144]

Fuck thank you, so many people try to make this so complicated

[@mayur9876]

Litt 👍👍

[@deeplizard]

Love it! 🔥

[@mysticmoose]

nice

[@deeplizard]

Thanks mm!

[@davidmurphy563]

Well, that's nice and simple so far. Tensors are multi-dimensional arrays. I bet you we're going to do matrix transforms - ah-hem, tensor transforms - on them. The thing is, if you want to transform space in such a way that you separate inputs into an nd array into say a normalised scalar value - ah-hem zero rank tensor - then you're not going to do that with just a linear transformation in anything but the simplest cases. You're going to have to, stretch then twist like with dough. With the linear, it would make sense to do an Eigenvector transform. But non-linear... God knows. I guess the point is that we don't know, and we just back prop / randomly adjust / hill climb / monty carlo guess our way to a better matrix transform. Ah-hem. Tensor transform.

[@therealmagicmonti]

I'm a physicist and I'm crying plz don't call this a tensor, what you mean just a multi-dimensional array or multilinear map. A tensor is meaningless if there's no defined basis, like a vector is meaningless without a basis. Maybe you should have a look at for example metrictensor fieldtensor or riccitensor. I just wondered what computer scientist mean about tensors because it's actually a concept from physics.

[@oguzbalasar9391]

I love you

[@XzionChughtai]

now i am confused because you made it sound like database Cube

[@ChargedPulsar]

So basically, it's a gimmick techy word for an (N) dimensional array. But noo, they wouldn't use something that existed before, they needed to be cool and invent some new name, which contradicts even the already meaning of "Tensor" for what it really is.

[@georgeigwegbe7258]

Hello World, it is PyTorch.

[@englishforfunandcompetitio248]

Rubbish

[@taasinbinhossainalvi9173]

Are you physicist?😂

[@LokeKS]

Corrupted use from physics

[@LokeKS]

See you in the vlog!