Initialize the weights.

Function to initialize the submodels of the model.

Iterates over all layers of the model.

Retrieves a list of all publicly available adapters on AdapterHub.ml or on huggingface.co. Args: model_name (str, optional): If spec

Loads a given adapter fusion configuration specifier into a full AdapterFusionConfig instance. Args: config (Union[dict,

Loads an adapter setup from the local file system or a remote location. Args: adapter_setup_name_or_path (str): can be either:

Loads the weights of a given adapter from a state dict into the model. Args: state_dict (dict): The state dict from whic

Loads the weights of a given AdapterFusion module from a state dict into the model. Args: state_dict (dict): The state d

Merges the weights of the given LoRA module with the Transformer weights as described in the paper. Args: name (str): Lo

Pads all key & value states to the longest prefix length in the current batch. This is required e.g. for stacked prefix tunings.

Performs computations after the forward pass of the adapter block(s). This e.g. includes applying the residual connection and layer n

Post-processes the predictions of a question-answering model with beam search to convert them to answers that are substrings of the original

Retrieves the hidden_states, query (for Fusion), and residual connection according to the set configuration. Args: adapt

Called before saving the adapters to disk.