About xt_next = at_next.sqrt() * x0_t + c1 * torch.randn_like(x0_t) + c2 * et

vvictoryuki / FreeDoM

[ICCV 2023] Official PyTorch implementation for the paper "FreeDoM: Training-Free Energy-Guided Conditional Diffusion Model"

260 stars 9 forks source link

About xt_next = at_next.sqrt() * x0_t + c1 * torch.randn_like(x0_t) + c2 * et #20

Closed SerJamie closed 3 weeks ago

SerJamie commented 3 weeks ago

Thank you for your wonderful work! However, I have a small question when corresponding the paper's formula with the code. How can I derive the formula xt_next = at_next.sqrt() x0_t + c1 torch.randn_like(x0_t) + c2 * et from the algorithm in the paper? Thank you!