Participare lunară
pe lunăUn loc recurent în conținutul meu. Abonații intră într-o coadă — sau într-o tragere la sorți când sunt mai mulți abonați decât sloturi — și apar în videoclipul în sine, în credite sau în descriere.
pe lună
Preț la cerere
Următorul
But how do AI images and videos actually work? | Guest video by Welch Labs 38 min 
SEXTANT 2 min 
Are Beavers Actually Good at Dams? 20 min 
Humanity’s Real Plan to Stop This Asteroid (Feat. Mark Rober) 22 min 
Игра Без Сохранений (Мото VLOG) 44 min 
Her Eyes Were Green 7 min 
I Tested NASA's New Space Suit (Ft. Axiom Space) 26 min 
NOT TODAY 3 min 
Rain 3 min 
The Real Reason We Should Revive Extinct Animals 16 min 
They Created A Skin Cancer Vaccine 1 min 
People Are Very Mad at Sony for This 13 min 
What’s Happened?? #secretgarage 2 min 
The vulnpocalypse might not be so bad after all 34 min 
The Napoleonic Wars - OverSimplified (Part 1) 30 min 
Magnets+Skateboard=Hoverboard 2 min 
Replacing Yourself with AI is Stupid 14 min 
There's No Way People Actually Like This Movie 26 min 
Why Fast Food Got So Expensive 20 min 
The Hidden Engineering Behind the Falkirk Wheel 16 min 
Name That Infrastructure (Ep. 3) #shorts 1 min 
Name That Infrastructure (Ep. 8) #shorts 1 min 
Lessons From Training Composer At Cursor And Building Meta/Nvidia Compute Clusters | YC Paper Club 77 min 
Name That Infrastructure (Ep. 9) #shorts 1 min 
Man Turns an Empty Space into a Dream Garage | @SmartEasyDIYer 
The Los Angeles Aqueduct is Wild 23 min 
The Hidden Engineering of Floating Bridges 18 min 
Why Roads Get Washboards 18 min 
How Long Does Sperm REALLY Live? #science #scishow #STEM #hankgreen 1 min 
Sneak peek at the making of a TED-Ed animation #shorts 1 min 
The incredible engineering of Venice - Stephanie H. Smith 7 min 
My Favorite Twilight Knockoff 101 min 
When did we start using passports? - Kristin Surak 7 min 
We Took Our Food Delivery Man on His First Vacation 46 min 
Living by the Electric Power Lines/Centers 13 min 
We’ve just launched *five* new channels! #shorts 1 min 
The Grid That Doubles the Strength of the Ground 19 min 
3 Million! 1 min 
LSD May Lead to a New Kind of Medicine 25 min 
SciShow is heading into the field TOMORROW with the all new Field Trips mini series. 1 min 
An Engineer's Perspective on the Texas Floods 24 min 
What toys have kids played with throughout history? 6 min 
Let's Go on a Field Trip! 3 min 
Why avalanches are most deadly when they stop - Simon Trautman 6 min 
This Glove Punches For You! (COLD GAS ROCKET FIST) 20 min 
1955 vs 2025, who actually had it better? 25 min 
Peppa's FUTURISTIC Mirror #PeppaPig #Shorts 1 min 
Weather CHAOS on the Family Lake Trip! #PeppaPig #Shorts 1 min 
Why Only 1% of All Possible Flights Are Direct 12 min 
Should you rinse your dishes before placing them in the dishwasher? - Rachel Yang 6 min 
Rate our gaming setup 🎮 #gta6 #gaming #desksetup 1 min But how do AI images and videos actually work? | Guest video by Welch Labs
2025 38 min
Diffusion models, CLIP, and the math of turning text into images Welch Labs Book: https://www.welchlabs.com/resources/imaginary-numbers-book Sections 0:00 - Intro 3:37 - CLIP 6:25 - Shared Embedding Space 8:16 - Diffusion Models & DDPM 11:44 - Learning Vector Fields 22:00 - DDIM 25:25 - Dall E 2 26:37 - Conditioning 30:02 - Guidance 33:39 - Negative Prompts 34:27 - Outro 35:32 - About guest videos Special Thanks to: Jonathan Ho - Jonathan is the Author of the DDPM paper and the Classifier Free Guidance Paper. https://arxiv.org/pdf/2006.11239 https://arxiv.org/pdf/2207.12598 Preetum Nakkiran - Preetum has an excellent introductory diffusion tutorial: https://arxiv.org/pdf/2406.08929 Chenyang Yuan - Many of the animations in this video were implemented using manim and Chenyang’s smalldiffusion library: https://github.com/yuanchenyang/smalldiffusion Cheyang also has a terrific tutorial and MIT course on diffusion models https://www.chenyang.co/diffusion.html https://www.practical-diffusion.org/ Other References All of Sander Dieleman’s diffusion blog posts are fantastic: https://sander.ai/ CLIP Paper: https://arxiv.org/pdf/2103.00020 DDIM Paper: https://arxiv.org/pdf/2010.02502 Score-Based Generative Modeling: https://arxiv.org/pdf/2011.13456 Wan2.1: https://github.com/Wan-Video/Wan2.1 Stable Diffusion: https://huggingface.co/stabilityai/stable-diffusion-2 Midjourney: https://www.midjourney.com/ Veo: https://deepmind.google/models/veo/ DallE 2 paper: https://cdn.openai.com/papers/dall-e-2.pdf Code for this video: https://github.com/stephencwelch/manim_videos/tree/master/_2025/sora Written by: Stephen Welch, with very helpful feedback from Grant Sanderson Produced by: Stephen Welch, Sam Baskin, and Pranav Gundu Technical Notes The noise videos in the opening have been passed through a VAE (actually, diffusion process happens in a compressed “latent” space), which acts very much like a video compressor - this is why the noise videos don’t look like pure salt and pepper. 6:15 CLIP: Although directly minimizing cosine similarity would push our vectors 180 degrees apart on a single batch, overall in practice, we need CLIP to maximize the uniformity of concepts over the hypersphere it's operating on. For this reason, we animated these vectors as orthogonal-ish. See: https://proceedings.mlr.press/v119/wang20k/wang20k.pdf Per Chenyang Yuan: at 10:15, the blurry image that results when removing random noise in DDPM is probably due to a mismatch in noise levels when calling the denoiser. When the denoiser is called on x_{t-1} during DDPM sampling, it is expected to have a certain noise level (let's call it sigma_{t-1}). If you generate x_{t-1} from x_t without adding noise, then the noise present in x_{t-1} is always smaller than sigma_{t-1}. This causes the denoiser to remove too much noise, thus pointing towards the mean of the dataset. The text conditioning input to stable diffusion is not the 512-dim text embedding vector, but the output of the layer before that, [with dimension 77x512](https://stackoverflow.com/a/79243065) For the vectors at 31:40 - Some implementations use f(x, t, cat) + alpha(f(x, t, cat) - f(x, t)), and some that do f(x, t) + alpha(f(x, t, cat) - f(x, t)), where an alpha value of 1 corresponds to no guidance. I chose the second format here to keep things simpler. At 30:30, the unconditional t=1 vector field looks a bit different from what it did at the 17:15 mark. This is the result of different models trained for different parts of the video, and likely a result of different random initializations. Premium Beat Music ID: EEDYZ3FP44YX8OWT
Țară
Statele Unite ale Americii
Lucrează cu acest creator
Pachete publicate de creatorul din spatele acestui videoclip.
Acest creator nu și-a stabilit încă tarifele. Trimite o cerere și o vom transmite — prețul este agreat înainte de a se percepe orice.
Plasare garantată — videoclip nou
unicăO apariție garantată de cel puțin 10 secunde în următorul meu videoclip. Prețul depinde de locul inserției în durata de rulare.
- Primele 1–3 minute
- Mijlocul videoclipului
- Sfârșitul videoclipului
Preț la cerere
- Primele 1–3 minute
- Mijlocul videoclipului
- Sfârșitul videoclipului
Plasare garantată — Short
unicăO apariție garantată de cel puțin 10 secunde într-un Short tăiat dintr-un videoclip existent al meu.
unică
Preț la cerere