Implement in subclasses of [`PreTrainedModel`] for custom behavior to prepare inputs in the generate method.

Usage: @profile("my_func") def my_func(): code

projects hidden states correctly to key/query states

Prunes heads of the model. heads_to_prune: dict of {layer_num: list of heads to prune in this layer}

OpenCV reads images as BGR, Pillow saves them as RGB. Work around this incompatibility to avoid colour inversions.

Register a new configuration in this mapping.

Register a new configuration for this class. Args: model_type (`str`): The model type like "bert" or "gpt".

Register a new tokenizer in this mapping. Args: config_class ([`PretrainedConfig`]): The configuration

Register a new model for this class. Args: config_class ([`PretrainedConfig`]): The configuration corres

sample.keys:question:str,context:stc, answer:[],idx:int,ans_span:[]

_summary_ Args: model (_type_): _description_ tokens (Tensor): [1, seq_len] attention_mask (Tensor): [1, 1, seq_len, seq_

Save the tokenizer vocabulary to a directory or file.

Save the tokenizer vocabulary to a directory or file.

Scale the parameters of a module and return it.

Set logging threshold on current logger.

Set model parallel rank.

Set the model parallel size