This is an extended degradation model by combining the degradation models of BSRGAN and Real-ESRGAN ---------- img: HXWXC, [0, 1], it

This is the degradation model of BSRGAN from the paper "Designing a Practical Degradation Model for Deep Blind Image Super-Resolution" --

This is the degradation model of BSRGAN from the paper "Designing a Practical Degradation Model for Deep Blind Image Super-Resolution" --

A util function for plotting a grid of images. Args: images: (N, H, W, 4) array of RGBA images rows: number of rows in the gr

Take a list of samples (as dictionary) and create a batch, preserving the keys. If `tensors` is True, `ndarray` objects are combined into ten

Overwrite model.train with this function to make sure train/eval mode does not change anymore.

Overwrite model.train with this function to make sure train/eval mode does not change anymore.

Overwrite model.train with this function to make sure train/eval mode does not change anymore.

bicubic downsampling + blur Args: x: HxWxC image, [0, 1] k: hxw, double sf: down-scale factor Return: downsam

bicubic downsampling + blur Args: x: HxWxC image, [0, 1] k: hxw, double sf: down-scale factor Return: downsam

Evaluate if new feature vectors are at the manifold.

extract depth maps from the density volume :param con_volume: [1, 1+C, dX, dY, dZ] can by con_volume or sdf_volume :param c2

Update: data (dict): { 'image0': (torch.Tensor): (N, 1, H, W) 'image1': (torch.Tensor): (N, 1, H

Args: feat0 (torch.Tensor): [M, WW, C] feat1 (torch.Tensor): [M, WW, C] data (dict) Update:

Args: x: [N, C, H, W]

Args: feat0 (torch.Tensor): [N, L, C] feat1 (torch.Tensor): [N, S, C] data (dict) mask_c0 (to

Multi-Head linear attention proposed in "Transformers are RNNs" Args: queries: [N, L, H, D] keys: [N, S, H, D]

Multi-head scaled dot-product attention, a.k.a full attention. Args: queries: [N, L, H, D] keys: [N, S, H, D]

Args: x (torch.Tensor): [N, L, C] source (torch.Tensor): [N, S, C] x_mask (torch.Tensor): [N, L] (optiona

Args: feat0 (torch.Tensor): [N, L, C] feat1 (torch.Tensor): [N, S, C] mask0 (torch.Tensor): [N, L] (optio

:param x: sparse tensor :return: sparse tensor

:param geometry_feat: geometry features indicates sdf [n_rays, n_samples, n_feat] :param rgb_feat: rgbs and image features [n_views,

Embeds x to (x, sin(2^k x), cos(2^k x), ...) Different from the paper, "x" is also in the output See https://github.com/bmild

:param position: can be 3d coord or interpolated volume latent :param normals: :param view_dirs: :param feature_vect

:param img_pred: [Npx, nviews, npatch, c] :param img_gt: [Npx, npatch, c] :return:

:param pred: [N_pts, 3] :param gt: [N_pts, 3] :param weight: [N_pts] :param mask: [N_pts] :return:

:param pred: [N_pts, Npx, 3] :param gt: [N_pts, Npx, 3] :param weight: [N_pts] :param mask: [N_pts] :return:

Computes the smoothness loss for a disparity image The color image is used for edge-aware smoothness :param disp: [B, 1, H, W