Loads a model from config and a ckpt if config is a path will use omegaconf to load

Loads a glb model into the scene.

Make a concat dataset from multiple folders Don't suport captions yet If paths is a list, that's ok, if it's a Dict interpret it as: k=fo

Mask borders with value Args: m (torch.Tensor): [N, H0, W0, H1, W1] b (int) v (m.dtype)

Take the mean over all non-batch dimensions.

source: https://github.com/openai/guided-diffusion/blob/27c20a8fab9cb472df5d6bdd6c8d11c8f430b924/guided_diffusion/losses.py#L12 Compute the K

Draw a bounding box annotation for an image.

Download the given URL and return a binary-mode file object to access the data.

:param pts: [N_rays, n_samples, 3] :param uv : [N_rays, 2] normalized in (-1, 1) :param normals: [N_rays, n_samples, 3] Th

:param pts: [N_rays, n_samples, 3] :param imgs: [N_views, 3, H, W] :param intrinsics: [N_views, 4, 4] :param c2ws: [

Get the distribution q(x_t | x_0). :param x_start: the [N x C x ...] tensor of noiseless inputs. :param t: the number of diff

Computes the percentage of pixels for which the ratio of the two depth maps is less than a given threshold. Takes preprocessed depths (no nan

Sample the ray in a given range and run secant on rays which have sign transition

Applies a rigid transform (c2w) to an (N, 3) pointcloud.

Saves rendered images of the object in the scene.

Runs the secant method for interval [z_low, z_high] for n_secant_steps

convert the sparse volume into dense volume to enable trilinear sampling to save GPU memory; :param coords: [num_pts, 3]

Run sphere tracing algorithm for max iterations from both sides of unit sphere intersection

Update: data (dict): { "conf_matrix_gt": [N, hw0, hw1], 'spv_b_ids': [M] 'spv_i_ids': [M]

Update: data (dict):{ "expec_f_gt": [M, 2]}