Run the speech decoder process. Parameters: - hidden (torch.Tensor): The output for embedding layer of the language model.

Disable the redundant torch default initialization to accelerate model creation.

Repeat module N times. :param int N: repeat time :param function fn: function to generate module :return: repeated modules :rtype: Mu

Add TDNN common arguments.

Create mask for subsequent steps (size, size). This mask is used only in decoder which works in an auto-regressive mode. This means the curre

Generates the model's next output based on the current input and state. Parameters: - inputs: The input tensor containing th

Merge `vision_embeddings` into `input_embeds` by overwriting the positions in `input_embeds` corresponding to placeholder image token

Add Subsampling common arguments.

Add Encoder common arguments.

Apply optional mask for encoder. Args: xs (torch.Tensor): padded input, (B, L, D), L for max length mask (torch.Tensor): mask fo