Integrals 300X Faster in Python (DON'T use SciPy)

CHECK OUT MY NEW UDEMY COURSE, NOW 90% OFF WITH THIS CODE:

Code:

Links to the torchquad library:

0:00 Introduction
1:14 torchquad webpages
2:25 Basic Usage
10:11 Physics Example Derivation
21:07 Physics Example Integral Solving

[TheUndergraduateQuant]

Love this content! Hope to see more of this and ODE, SDE, etc…

[ntwede]

9:06
I Lol'd when I saw "simp" as a var name :)

But seriously, thank you for these vids. Your explanations, and example code repo's are super helpful!

[lucasasth]

Can you make a video of this series discussing PDEs? Particularly about the finite difference method or any other numerical solution method. I think it would be really cool, and it's a fantastic application in the world of physics. Great series btw

[annapopczyk]

I'm really happy to see your channel grow so much. When I joined there was like not even 3k subs but look at you now! Congratulations 🥳

[dvir-ross]

Awesome content! Bought your course to support you 🙂

[chrisogonas]

Superb! That's very well illustrated. Thanks

[drewnotdrey]

Please create something like this for differential equations!! Maybe even use the differential equations used to simulate the double pendulum.

[ashishawasthi]

How on earth a content like this is so under-subscribed….😢(looks like everyone else is taking selfies with lions moments before getting mauled to death…)
Amazing work dear….

[tupui]

Array API support will soon make SciPy use transparently any backend you want.

[RalphDratman]

I am glad you can keep all those variables and units straight in your mind.
I never could, maybe because I have ADD. Unfortunately.
So, physics was not really an option for me after the first few quarters.
But anyway, you make a fine example of a careful physics student (or professor, maybe?).

[amzion]

Please what is your timer under your cells with the little check and the time showing ?

[valevan]

hey I don't know if you have ever done adaptive integration in python? I have been making a function in python which calculates a certain function in logarithmic grid (the grid where you have geometric series spacing from min to max in positive axis and reflect it in negative axis). but, it oscillates wildly near zero. I have tried to add more points in the area near zero but it still gives nothing significant that changes the result better. appreciate all kinds of help!

[calebwhales]

I hate to watch the video without trying myself, but I work with a instrumentation that doesn't give output as a function but rather as a table of values, more of a mapping function. So the xaxis would be say values from 70 to 320 degrees, with yaxis some value such that the result is a complicated waveform. I wonder if this approach would be useful to calculate the integral of that waveform in real time?

[kaouthartlili]

thank you so much for the video, when changing N in my simp() integration, the result changes, why?

[ralphwalters]

Will TorchQuad have access to the GPU cores in the Apple M3 processor ?

[mjbgcdzqsxfthn]

its too much for me, im gonna explode 🤯🤯

[RalphDratman]

So in the magnet-drop example you are comparing two different ways to compute a numerical integral, is that right?
You are not trying to find an analytical integral for that integrand. Correct?

[AntonioNoack]

The first comparison feels unfair, as we don't know how many steps numpy is taking. 37 seems very arbitrary.

[vidal]

37 points per dimension is a little too little, no? I would try using more than that to prove the point better.

[wildbirbs]

Hi! could you show us your parrot in the next video? it would be awesome and cute