Sets the discrete timesteps used for the diffusion chain (to be run before inference). Args: num_inference_steps (`int`)

Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion process from the learned m

Apply a function to each element of a batch.

Save a video to disk.

embed_dim: output dimension for each position pos: a list of positions to be encoded: size (M,) out: (M, D)

grid_size: int of the grid height and width return: pos_embed: [grid_size*grid_size, embed_dim] or [1+grid_size*grid_size, embed_dim] (w/ or

grid_size: int of the grid return: pos_embed: [grid_size, embed_dim] or [1+grid_size, embed_dim] (w/ or w/o cls_token)

r""" Set the attention processor to use. Args: processor (`AttnProcessor`): The attention processor to us

Args: x (torch.Tensor): of shape [B, N, C] Return: x (torch.Tensor): of shape [B, C_out, T, H, W]

embed_dim: output dimension for each position pos: a list of positions to be encoded: size (M,) out: (M, D)

LlamaRMSNorm is equivalent to T5LayerNorm

compatible with diffusers add_noise()

Returns: outputs: Tensor handle: Optional[Work], if overlap is True

embed_dim: output dimension for each position pos: a list of positions to be encoded: size (M,) out: (M, D)

Load a video from the given path and convert it to a PyTorch tensor.

Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles custom timesteps. Any kwargs

Ensures interchangeability with schedulers that need to scale the denoising model input depending on the current timestep. A

r""" Set the attention processor to use. Args: processor (`AttnProcessor`): The attention processor to us