Видео

Amazing video ✨

this is awesome

Why is grid search computationally more expensive - it depends on how many grid points there are - if your random search has more random hp picks, then the random search takes longer, no?

Thank you! That visualisation with explaining sample mean being closer made A LOT OF SENSE! Though I do not understand the degrees of freedom part and it always bothered me! Could you make a video on degrees of freedom as well (I can see how we get x3 from the other info but I do not get how it helps anything and how it is related to dividing 2?? Thanks a lot in advance!)

great video, thx!

thank you for the explantion!

great job, finally useful video to make this theorem simple and clear , keep working

But how do you get Q? How do you get K? How are the values of those vectors defined?

So in poor words. Oscillation in time domain gets converted into a sum of rotational frequencies in complex domain?

Fascinating stuff cheers mate

Video errors: - at 03:36 I should have written "low variance" on the screen for underfitting instead of "high variance". Thanks @nuktu for noticing this!

Full video at: ruclips.net/video/B9rhzg6_LLw/видео.html

Great explanation! Thankyou so much!

dada tank u

Where did you get the original explanation for this?

Why are the circles getting bigger as you minimise the loss?

Great explanation!

Learn more about the differences between L1 and L2 regularization here: ruclips.net/video/aBgMRXSqd04/видео.html

Question, if in 2:15 the feature map passes through a CNN of 3X3 , then h and w in the minute 3:05 wouldn't be of the height and width of the feature map less 2 ?(3-1 both due to the fact that is 3X3 the CNN) or may be a padding of 1 should be applied first to the feature map ?

This is probably the best explanation of this concept on RUclips. Every other video on standard deviation glosses over why we use Bessel’s correction.

Can you elaborate more on how the degrees of freedom are connected to the bias? Why is it that we need to consider degrees of freedom in the sample metrics but not in the population metrics?

Sometimes the market needs a good laugh, and $LLM delivers. But jokes aside, it’s launching on SOLANA and hitting BingX tomorrow. I’ll admit it-I’m curious to see where this goes.

recently i discovered this channel and i have to say your explanations are so simple its very good to understand and i want to thank you

Omg, thank you for the explanation! I was eagerly searching for this one ❤

Arey bhai aur thode khatin words use kar le na simple chiz ko samjhane ke liye

Excellent

Lovely video. Great explanation!!!🎉🎉

Why do I get a different mean from your sample of [173,173,172,161] in your example at 2:00? I get a mean of 169.75 not 170.38.

nice video dude clearly explaining the concept, thx

-1 is not going upward, to the previous hidden state, but to the right, toward the point wise multiplication of zt and h(tilde)t

At 03:36 you say "low variance" for underfitting models, but the screen says "high variance" on the right hand side box. I assume that should say "low variance" too?

Full video link: ruclips.net/video/8ZYLWipgPT0/видео.html

One question, for the total gradient for we need to calculate for o1,o2 too and add them together right?

A more theoretical overview of these issues are presented in my video about the bias-variance trade-off: ruclips.net/video/5mbX6ITznHk/видео.html

Must Love them Hungarians!

Thank you so much for explaining this topic in such an easy way.

preparing for a job interview and im handling object detection, this is definitely a must for everyone interested in learning more about this topic.

So, BART was made “outdated” by which new technology? What is the best tool for NLP tasks?

Thanks for the clear explanation!

how we get shared key? bro

This was a really good explanation of the Fourier Teansform.

Correction -- the Gaussian distribution is NOT on the *perpendicular distance* of the input-label tuple from the mean line, but instead the *vertical distance* (along the label dimension).

This is a very nice crisp content. Thank you.

The best ML channel

Amazing content my friend, keep going !

Waaaejtg5f e rf

Man, it's literally the first time I ever comment a video on Yt. I don't have much time to prepare for the exam, but your delievery and explanations are so clear. Be blessed and keep on going!!

Thanks for this! very useful for revision when preparing for a test or interview.

You didn't explain anything. What is the intuition behind q,k,v ? Just rewrote the formula