Extract values from a 1-D numpy array for a batch of indices. :param arr: the 1-D numpy array. :param timesteps: a tensor of indices int

Updates the parameter list before the forward prop so that if `params` is None or doesn't have a certain key, the module uses the def

Load the mesh from a .npz file.

Take the mean over all non-batch dimensions.

Create sinusoidal timestep embeddings. :param timesteps: a 1-D Tensor of N indices, one per batch element. These may be

Input: points: input points data, [B, N, C] idx: sample index data, [B, S] Return: new_points:, indexed points data,

Creates a copy of self where some or all values are transformed by map_fn. :param should_map: If provided, only those values

Evaluate a function without caching intermediate activations, allowing for reduced memory at the expense of extra compute in the backward pas

The ranges at which rays intersect with either one, or both, or none of the self and another are merged together.

Perform volumetric rendering over a partition of possible t's in the union of rendering volumes (written below with some abuse of notations)

Instantiate the encoder with information about the renderer's input parameters. This information can be used to create output layers

Appends dimensions to the end of a tensor until it has target_dims dimensions.

Embed images into latent grids. :param xs: an iterable of images to embed. :return: a tensor of shape [N x C x L], where L =

Converts a denoiser output to a Karras ODE derivative.

Required: d_input, d_hidden, d_output Optional: act_name, bias

Create a beta schedule that discretizes the given alpha_t_bar function, which defines the cumulative product of (1-beta) over time from t = [

Apply the model to get p(x_{t-1} | x_t), as well as a prediction of the initial x, x_0. :param model: the model, which takes

Concatenate x and its positional encodings, following NeRF. Reference: https://arxiv.org/pdf/2210.04628.pdf

Compute the mean and variance of the diffusion posterior: q(x_{t-1} | x_t, x_0)

:param position: [batch_size x ... x 3] :param params: Meta parameters :param options: Optional hyperparameters :retu

Get a term for the variational lower-bound. The resulting units are bits (rather than nats, as one might expect). This allow

Cast a batch of rays onto a mesh.

Some values may be missing, but the dictionary structures must be the same. :param combine_fn: a (possibly non-commutative)

:param xs: N images, stored as numpy arrays, tensors, or PIL images. :return: an [N x D] tensor of features.

Embed text prompts as an [N x D] tensor.

Load the given channels from the view at the given index. :return: a tuple (camera_view, data), where data is a float array of

Diffuse the data for a given number of diffusion steps. In other words, sample from q(x_t | x_0). :param x_start: the initi

Save the mesh to a .npz file.

Returns a [batch x num_views x 3 x size x size] tensor in the range [-1, 1].

A fast approximation of the cumulative distribution function of the standard normal.

For every (x, y) coordinate in a rendered image, compute the ray of the corresponding pixel. :param coords: an [N x 2] integ

Center crops an image.

Combines the integration results of `self` over [t0, t1] and `cur` over [t1, t2] to produce a new set of results over [t0, t2] by

Compute what the p_mean_variance output would have been, should the model's score function be conditioned by cond_fn. See co

Returns a tuple of (rays, z_directions) where - rays: [batch, num_views, height, width, 2, 3] tensor of camera information.

Returns a [batch x num_views x 1 x size x size] tensor in the range [-1, 1].

:return: [batch_size, num_views, n_patches, width]

Input: xyz: pointcloud data, [B, N, 3] npoint: number of samples Return: centroids: sampled pointcloud index, [B, npo

Sample a subset of the point cloud that is evenly distributed in space. First, a random point is selected. Then each successive poin

Download the file at the given URL into a local file and return the path. If cache_dir is specified, it will be used to download the files.

:param position: [batch_size x ... x 3] :param params: Meta parameters :param options: Optional hyperparameters