x : torch.LongTensor, shape = (batch_size, <= n_ctx) the text tokens xa : torch.Tensor, shape = (batch_size, n_mels, n_au

r""" Args: sample (`torch.FloatTensor`): (batch, channel, height, width) noisy inputs tensor timestep (`torch.FloatTen

Get all the stored features as PyTorch Tensor. Returns: Concatenation of the stored features.

Get the mean and covariance of the stored features. Returns: Mean and covariance of the stored features.

Sample a random clip starting index from the video. Args: idx: Index of the video.

Get sliced features based on a given index :param feature_array: :param start_idx: the start index of the feature :pa

Check if the maximum number of samples is reached. Returns: True if the storage is full, False otherwise.

Returns the token id corresponding to the value of the `language` field

Returns the list of tokens to suppress in order to avoid any speaker tags or non-speech annotations, to prevent sampling texts that a

Compute left and right padding

Run a single prediction on the model

Replace any other markers, symbols, punctuations with a space, keeping diacritics

Load the model into memory to make running multiple predictions efficient