I'm rewatching the video, and it looks like many effects & transitions are missing. My editor kept crashing so some effects were probably lost then. Guess I shouldn't be recording in 4K constantly, my 2013 macbook can't handle it. At the very least, the most important part, i.e. the main content, is intact. Hope you still enjoy the video!
I'm currently a graduate student at the University of Southern California. Love reading up on trending AI. Also like writing stuff on Quora too: www.quora.com/profile/Ajay-Halthor Glad to know some people are interested :P
Actually your voice and expressions are a clue into your intelligence 💪🏽 and you uniquely make the whole think more interesting . Soooo good job friend 🤩
It would be great if you could actually run a small sample yourself for everything you teach. This makes it more authentic and also will get you much bigger reach!
Oh shit! I am actually understanding this. I hope you were the face of AI youtubers rather than a "certain someone" who committed multiple IP thefts and business frauds. Great job man! Subscribed.
@CodeEmporium this was a great explanation of the paper and helped me solve few of the doubts i had regarding the paper. Did you use any slides for the presentation? If yes can you please put a shareable links for the slides in the description. Thanks
I, too, have this question. Are we left to implement their paper on our own? Curious because this seems like a fun experiment to use with the dataset as well as other unseen speakers such as friends.
Good day, I would to ask if there is a source code that can be converted to java to use as a thesis of my friend for autism. It will be a great help if you could send me an email. Thanks
If you learn undergrad statistics and ML basics that's actually pretty easy to understand. The visualization from around 3:15 also really helps. Code implementation however won't help since you have to understand the math first, then implement it with code, not the other way around.
3:07 Quite misleading for me. I can only think of GMMs for analogy here. For 2 speaker identification, we need 2 GMMs (each GMM has its own number of mixture components) I am comfortable with "n distributions used for n classes". However since you used a single distribution for n classes, it is quite misleading for me.
Apparently there are two ways. Both work it seems. So for voice cloning, the generative modelling approach here is to go with a single distribution with each bump associated with a different speaker?
The paper shared in the description, has no mention of "MFCC". The distribution's X-Y plane is supposed to be representing some feature vector, like MFCC.
when they come out with a affordable PC based software that is not subject to the cloud or pay as you go network.. Then I am onboard. I want a program I can own, and install it to my PC... a stand alone software. If anyone knows of a program that fits that bill.. Please reply I would really appreciate it. since it is 2021 now...
Thanks for sharing the video! Let me tell you something important: God loves you! God loves us ! He he is no respecter of persons! because God wants us all to be saved! and let's go to the knowledge of the "Truth" God wants to save us from eternal damnation, which we all deserve because of sin, who entered the world through Adam, But God's Love was so great for us, that he sent his only Son (Jesus Christ), gave him to this world, to die and rise again for All of us! so that everyone who believes in Jesus Christ, does not go to eternal punishment, but has “Eternal Life! " Jesus Christ came to call sinners to “Repentance! " and, whoever believes in Him, has "Eternal Life! “But whoever refuses to believe in Jesus Christ, the wrath of God remains upon him. All of us who believe in "The Only Savior of the world! " The Only Mediator between God and us! The Lord Jesus Christ! " All of us who trust in Him; we have "Eternal Life! " God loves you! God loves us! Only to Him be the Glory Forever! Amen! Biblical Source: Acts 10: 34-35 / 1 Timothy 2: 3-6 / Romans 5:12 / John 3:16 / Luke 5:32 / Matthew 16:21 / 1 Corinthians 15: 20-22 / John 3:36 / Romans 5:18 / Acts 4:12 / John 6:47 /
![The moment we stopped understanding AI [AlexNet]](http://i.ytimg.com/vi/UZDiGooFs54/mqdefault.jpg)
![The moment we stopped understanding AI [AlexNet]](/img/tr.png)
This channel is waaaay undervalued
I'm rewatching the video, and it looks like many effects & transitions are missing. My editor kept crashing so some effects were probably lost then. Guess I shouldn't be recording in 4K constantly, my 2013 macbook can't handle it. At the very least, the most important part, i.e. the main content, is intact. Hope you still enjoy the video!
CodeEmporium are you studying? Would like to know more about you
I'm currently a graduate student at the University of Southern California. Love reading up on trending AI. Also like writing stuff on Quora too: www.quora.com/profile/Ajay-Halthor
Glad to know some people are interested :P
mac sucks, but i appreciate your work. thanks!
Extremely well done video. Thank you.
Dang,this hurts my brains
I will try and wrap this around my head somehow..
Thanks for the video!
Nice, continue what you're doing. I like how you go in depth and take things step by step without jumping around.
Thanks! Kinda what I was going for.
Actually your voice and expressions are a clue into your intelligence 💪🏽 and you uniquely make the whole think more interesting . Soooo good job friend 🤩
It would be great if you could actually run a small sample yourself for everything you teach. This makes it more authentic and also will get you much bigger reach!
Yep a tutorial side of it
It’s pretty scary. Imagine what this could do in the wrong hands!
nice video dude. Keep them coming!
Will do Mark! Thanks!
Thank you
● But what is the loss function
● And what kind of activation function have used
And what kind of NN it's depending on
Amazing video, this was well explained! Thanks!
Thanks Baidu!
An update would be cool!
Oh shit! I am actually understanding this. I hope you were the face of AI youtubers rather than a "certain someone" who committed multiple IP thefts and business frauds. Great job man! Subscribed.
Finally,one good thing Baidu did.
the main issue here that's we use pretrained modele so the modele generate only voices in english but ty a lot
1:48 *_P U D D I N G_*
@CodeEmporium this was a great explanation of the paper and helped me solve few of the doubts i had regarding the paper. Did you use any slides for the presentation? If yes can you please put a shareable links for the slides in the description.
Thanks
So what's the the best opensource method to start playing around with implementation?
I, too, have this question.
Are we left to implement their paper on our own? Curious because this seems like a fun experiment to use with the dataset as well as other unseen speakers such as friends.
did anyone find an implementation?
Good day, I would to ask if there is a source code that can be converted to java to use as a thesis of my friend for autism. It will be a great help if you could send me an email. Thanks
That guy looks like he was created by an ai. No human has those eyes and hand gestures.
They're on to me dammit
7:02 this is incomprehensible. Why don't you just show how stuff is done in code instead so it's accually possible to understand how this is made.
If you learn undergrad statistics and ML basics that's actually pretty easy to understand. The visualization from around 3:15 also really helps. Code implementation however won't help since you have to understand the math first, then implement it with code, not the other way around.
um why would that be cool? , how is this going to result in anything but extreme negative outcomes? like magpies and a shinny button
Can I clone Hindi voice using this model?
3:07 Quite misleading for me.
I can only think of GMMs for analogy here. For 2 speaker identification, we need 2 GMMs (each GMM has its own number of mixture components)
I am comfortable with "n distributions used for n classes". However since you used a single distribution for n classes, it is quite misleading for me.
Apparently there are two ways. Both work it seems.
So for voice cloning, the generative modelling approach here is to go with a single distribution with each bump associated with a different speaker?
The paper shared in the description, has no mention of "MFCC". The distribution's X-Y plane is supposed to be representing some feature vector, like MFCC.
Weird/possibly stupid question: could you generate English-speaking AI using data from someone not speaking English?
lmao try Google Translate. Set the conversion,let's say, from German to English. And instead of typing words in German,give English words as input.
th-cam.com/video/38ZXwJj6j8k/w-d-xo.html
when they come out with a affordable PC based software that is not subject to the cloud or pay as you go network..
Then I am onboard.
I want a program I can own, and install it to my PC... a stand alone software.
If anyone knows of a program that fits that bill..
Please reply I would really appreciate it. since it is 2021 now...
first indian talk english good ♥
bro, your eyes.... kind of scary
I'm the stuff of nightmares.
there is not code
I mean, it sounds extremely robotic.
Good video
Thanks!
hey codeemporium can we do this task with the python programming
@@MeNoOther please share the link
next time increase the volume of your voice cuz i got my head phones in aur still low
George Michael?
Speaking
Thanks for sharing the video! Let me tell you something important:
God loves you! God loves us !
He he is no respecter of persons! because God wants us all to be saved! and let's go to the knowledge of the "Truth"
God wants to save us from eternal damnation,
which we all deserve because of sin,
who entered the world through Adam,
But God's Love was so great for us, that he sent his only Son (Jesus Christ), gave him to this world, to die and rise again for All of us!
so that everyone who believes in Jesus Christ, does not go to eternal punishment, but has “Eternal Life! "
Jesus Christ came to call sinners to “Repentance! "
and, whoever believes in Him, has "Eternal Life! “But whoever refuses to believe in Jesus Christ, the wrath of God remains upon him.
All of us who believe in "The Only Savior of the world! "
The Only Mediator between God and us! The Lord Jesus Christ! "
All of us who trust in Him; we have "Eternal Life! "
God loves you! God loves us! Only to Him be the Glory Forever! Amen!
Biblical Source: Acts 10: 34-35 / 1 Timothy 2: 3-6 / Romans 5:12 /
John 3:16 / Luke 5:32 / Matthew 16:21 / 1 Corinthians 15: 20-22 / John 3:36 / Romans 5:18 / Acts 4:12 / John 6:47 /
and all the github links to the actual code so smolbrains like me can run it are 404