Causal 3d convolusion.

`FlowDPMSolverMultistepScheduler` is a fast dedicated high-order solver for diffusion ODEs. This model inherits from [`SchedulerMixin`] and [

`UniPCMultistepScheduler` is a training-free framework designed for the fast sampling of diffusion models. This model inherits from [`Schedu

Causal self-attention with a single head.

Adds gradient checkpointing to models that don't have native support. Ensures proper gradient flow.

Utility class to efficiently checkpoint model state during optimization.

Memory-efficient implementation of motion variance optimization during diffusion sampling. Uses manual gradient checkpointing for models with

XLMRobertaModel with no pooler and no LM head.

Wan diffusion backbone with improved GPU distribution