Args: x: HxWxC image, [0, 1] sf: down-scale factor Return: bicubicly downsampled LR image

Args: x: HxWxC image, [0, 1] sf: down-scale factor Return: bicubicly downsampled LR image

A counter for the `thop` package to count the operations in an attention operation. Meant to be used like: macs, params = tho

Return the data prediction model (with thresholding).

:param h: height :param w: width :return: normalized distance to image border, wtith min distance = 0 at border

Convert the continuous-time `t_continuous` (in [epsilon, T]) to the model input time. For discrete-time DPMs, we convert `t_continu

A piecewise linear function y = f(x), using xp and yp as keypoints. We implement f(x) in a differentiable way (i.e. applicable for autograd

Multistep DPM-Solver with the order `order` from time `t_prev_list[-1]` to time `t`. Args: x: A pytorch tensor. The in

Return the noise prediction model.

Restore the parameters stored with the `store` method. Useful to validate the model with EMA parameters without affecting the

shift pixel for super-resolution with different scale factors Args: x: WxHxC or WxH sf: scale factor upper_left: shift

shift pixel for super-resolution with different scale factors Args: x: WxHxC or WxH sf: scale factor upper_left: shift

Singlestep solver DPM-Solver-3 from time `s` to time `t`. Args: x: A pytorch tensor. The initial value at time `s`.

Save the current parameters for restoring later. Args: parameters: Iterable of `torch.nn.Parameter`; the parameters to b

Zero out the parameters of a module and return it.

USM sharpening. borrowed from real-ESRGAN Input image: I; Blurry image: B. 1. K = I + weight * (I - B) 2. Mask = 1 if abs(I - B) > thre

generate an anisotropic Gaussian kernel Args: ksize : e.g., 15, kernel size theta : [0, pi], rotation angle range l1

generate an anisotropic Gaussian kernel Args: ksize : e.g., 15, kernel size theta : [0, pi], rotation angle range l1

Appends dimensions to the end of a tensor until it has target_dims dimensions. From https://github.com/crowsonkb/k-diffusion/blob/master/k_diffus

Kai Zhang (github: https://github.com/cszn)

Create a 1D, 2D, or 3D average pooling module.

bgr version of rgb2ycbcr only_y: only return Y channel Input: uint8, [0, 255] float, [0, 1]

Compute the gradient of the classifier, i.e. nabla_{x} log p_t(cond | x_t).

Denoise at the final step, which is equivalent to solve the ODE from lambda_s to infty by first-order discretization.

The adaptive step size solver based on singlestep DPM-Solver. Args: x: A pytorch tensor. The initial value at time `t_

Get the order of each step for sampling by the singlestep DPM-Solver. We combine both DPM-Solver-1,2,3 to use all the function eval

This matches the implementation in Denoising Diffusion Probabilistic Models: From Fairseq. Build sinusoidal embeddings. This matc

imgs: list, N images of size WxHxC