The [`Transformer2DModel`] forward method. Args: hidden_states (`torch.LongTensor` of shape `(batch size, num latent pix

r""" The [`UNet2DConditionModel`] forward method. Args: sample (`torch.FloatTensor`): The noisy input ten

-- Texture Rendering vertices: [batch_size, V, 3], vertices in world space, for calculating normals, then shading transformed

Pytorch forward Args: x: input image (224x224) Returns: class logits

Input: shape_params: N X number of shape parameters expression_params: N X number of expression parameter

texcode: [batchsize, n_tex] texture: [bz, 3, 256, 256], range: 0-1

recon_x: generating images x: origin images mu: latent mean logvar: latent log variance

Function that converts angles from radians to degrees. See :class:`~torchgeometry.RadToDeg` for details. Args: tensor (Tensor): Tens

-- rendering colors: could be rgb color/ normals, etc colors: [bz, num of vertices, 3]

-- rendering depth

computes ring loss for ring_outputs before FLAME decoder Inputs: ring_outputs = a list containing N streams of the ring;

An api for running deca given an image path

image: 0-255, uint8, rgb, [h, w, 3] return: detected box

vertices: [nv, 3], tensor texture: [3, h, w], tensor

Disables custom attention processors and sets the default attention implementation.

assume: shading should be smooth ref: Lifting AutoEncoders: Unsupervised Learning of a Fully-Disentangled 3D Morphable Model using Deep Non-R

regularize lighting: assume lights close to white