Linear layer with column parallelism. The linear layer is defined as Y = XA + b. A is parallelized along its second dimension as A = [A_1, ..

Linear layer with row parallelism. The linear layer is defined as Y = XA + b. A is parallelized along its first dimension and X along its sec

A class that keeps all conversation history.

Diffusion model with a Transformer backbone.

Diffusion model with a Transformer backbone.

Track a series of values and provide access to smoothed values over a window or the global series average.

Multi-head attention module.

Utilities for training and sampling diffusion models. Original ported from this codebase: https://github.com/hojonathanho/diffusion/blob/

Utilities for training and sampling diffusion models. Original ported from this codebase: https://github.com/hojonathanho/diffusion/blob/

The final layer of DiT.

The final layer of DiT.

r"""Embeds class labels into vector representations. Also handles label dropout for classifier-free guidance.

Embeds scalar timesteps into vector representations.

Embeds scalar timesteps into vector representations.

r"""A helper class to load a tensor parallel sharded tensor and track the loading status (i.e., check if a tensor is loaded with consistent tensor

A diffusion process which can skip steps in a base diffusion process. :param use_timesteps: a collection (sequence or set) of timesteps from

A diffusion process which can skip steps in a base diffusion process. :param use_timesteps: a collection (sequence or set) of timesteps from