Read data from a given ``filepath`` with 'rb' mode. Note: There are two types of return values for ``get``, one is ``bytes``

Check whether a file path exists. Args: filepath (str or Path): Path to be checked whether exists. Returns:

Print a message only if VERBOSE environment variable is set to 1. Args: message: The message to print

Print a message regardless of VERBOSE setting. Use for important information, summaries, and critical messages. Args: message: T

Synchronize all processes in distributed mode.

Broadcast the tensor from the minimum rank in the specified group(s).

Edit the data type of the condition. Args: data_type: The new data type. Returns: A new Text2WorldCondition

Converts the condition to a dictionary. Returns: Dictionary containing the condition's fields and values.

We hanlde two types of data_batch. One comes from a joint_dataloader where "dataset_name" can be used to differenciate image_batch and video_batch.

r"""Read function for reading the data stream. Args: amt (int, optional): Amount of data to read Returns: chun

Load a config file. Args: filename: absolute path or relative path w.r.t. the current working directory keys

Load network weights and optimizer states from a checkpoint in a single process. The priority of the checkpoint loading logic is: 1.

Return a file backend based on the prefix of uri or backend_args. Args: uri (str or Path): Uri to be parsed that contains the file prefix

Sets the video conditioning frames for video-to-video generation. This method creates a conditioning mask for the input video frames

Broadcast and split the input data and condition for model parallelism. Currently, we only support context parallelism.

Recursively instantiate objects defined in dictionaries by "_target_" and arguments. Args: cfg: a dict-like object with "_target

Translate a `filepath` to a string. Paths are of the form 'path/to/file' (path form) or '{protocol}://path/to/file' (URL form). Args

Convert a ``filepath`` to standard format of s3 oss. If the ``filepath`` is concatenated by ``os.path.join``, in a Windows environmen

Map ``filepath`` to a string path whose prefix will be replaced by :attr:`self.path_mapping`. Args: filepath (str or Path

Compute text embeddings for the given prompts.

Remove a file. Args: filepath (str, Path): Path to be removed.

Encodes text prompts into T5 embeddings. Args: prompts: A single text prompt or a list of text prompts. encoder: An optional Cosm

Convert quaternion(s) to rotation matrix/matrices. Args: q (torch.Tensor [...,4]): Quaternion(s) (x, y, z, w). Returns:

Return checkpoint config for UUID.

Write data to a given ``filepath`` with 'wb' mode. Note: ``put`` should create a directory if the directory of ``filepath``

Concatenate outputs from different ranks in the checkpoint parallelism group. This function gathers tensors from all ranks in the checkpoint

Performs denoising on the input noise data, noise level, and condition Args: xt (torch.Tensor): The input noise data.

state : pixel range [-1, 1]

Return checkpoint path for S3 URI, HuggingFace URI, or local path. Supports: - S3 URIs: s3://bucket/path/to/checkpoint - HuggingFace URIs

Broadcasts and splits the condition across the checkpoint parallelism group. For most condition, such as Text2WorldCondition, we do not need s

Processes the provided data batch to generate two sets of outputs: conditioned and unconditioned. This method manipulates the dropout

Similar functionality as get_condition_uncondition But use negative prompts for unconditon

Create a "directory" in S3. Note: S3 doesn't have real directories, but we can create an empty object with a trailing slash

Save network weights, optimizer parameters, scheduler parameters to a checkpoint. Args: model (ImaginaireModel): The PyTorch mode

Make a copy of the input config that doesn't require S3 for checkpointing and I/O in the callbacks.

Upload a local file src to dst and return the destination file. Args: src (str or Path): A local file to be copied. d

Euler angle (ZYX) to rotation matrix.

Returns the compute capability of a given torch device if it's a CUDA device, otherwise returns 0. Args: device (torch.device | None

Import all modules from the specified package path recursively. This function is typically used in conjunction with Hydra to ensure that all

Normalize quaternion(s) to unit length. Args: q (torch.Tensor [...,4]): Input quaternion(s). eps (float): Small epsil

Save network weights, optimizer parameters, scheduler parameters to a checkpoint. Args: model (ImaginaireModel): The PyTorch mode

Get the file size in bytes for a given ``filepath``. Args: filepath (str or Path): Path to get file size in bytes. Retur

Broadcast model states from replicate mesh's rank 0.

Check whether a file path is a file. Args: filepath (str or Path): Path to be checked whether it is a file. Returns:

r""" Args: labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing

Normalizes video data in-place on a CUDA device to reduce data loading overhead. This function modifies the video data tensor within

Copy the file src to local dst and return the destination file. If dst specifies a directory, the file will be copied into dst using