Diffusion Models (DDPM & DDIM) - Easily explained!

[@jiananwang2681]

You helped me a lot! Thanks! Please keep going~

[@soroushmehraban]

@@jiananwang2681 Thanks for watching🙂

[@rafa_br34]

Awesome, I'd love to see a video about how high-fidelity VAEs work and how they're trained.

[@yipengsun8624]

clear!

[@soroushmehraban]

@@yipengsun8624 Thanks!

[@armanhatami5706]

awesome soroush. nice and clear explain

[@soroushmehraban]

@@armanhatami5706 Thanks 😃

[@ai1998]

you are great , please keep going

[@soroushmehraban]

Thanks Ahmed! Appreciate it.

[@GopalSharma-sf1zz]

Nice short explanation!

[@soroushmehraban]

Thanks!

[@oblivitus.]

Brilliant! thank you!!

[@soroushmehraban]

@@oblivitus. Thanks for watching

[@HassanHamidi-v8s]

Wow! great video. thanks a lot.

[@soroushmehraban]

@@HassanHamidi-v8s Thanks for watching it 🙂

[@abrarrahmanabir9588]

Best❤

[@vidaadelimosabeb6689]

Great video, keep it up!

[@soroushmehraban]

@@vidaadelimosabeb6689 Thanks😃

[@WaltonBoyd-c8g]

Ariane Gateway

[@moeinshariatnia59]

at 5:58 " ... the variance comes from the network" This is not right. In DDPM, the authors made it constant and then in later studies people started to make those learnable as well.

[@soroushmehraban]

@@moeinshariatnia59 In that formulation it is. Later in video I mentioned that it’s not necessary and model has to only predict the noise sampled from zero mean unit variance.

[@moeinshariatnia59]

@@soroushmehraban It's so funny you delete the comment instead of correcting the mistake :))

[@soroushmehraban]

@@moeinshariatnia59 I didn't delete anything. It wasn't a mistake. Sorry if my explanation was ambiguous.