Wrap T5 encoder to obtain a Fusion-in-Decoder model.

Wrap T5 encoder to obtain a Fusion-in-Decoder model.

Return a string representation of the item with its main attributes.

r"""Check if the index contains document content

r"""Check if the dir path exists and if there is content.

Collect all classes based on ```BaseMetric```.

Compute the log probabilities of the given labels under the given logits. Args: logits: Logits of the model (unnormalized). Shape

r""" Returns the default logging level.

r""" Returns the jinja template.

r""" Processes visual inputs for mllama because its image processor only accepts List[List[ImageInput]]. Returns: pixel_v

r""" Gets paligemma token type ids for computing loss. Returns: batch_token_type_ids: shape (batch_size, sequence_length)

r""" Prepares the tokenized dataset to perform AutoGPTQ. Do not use tensor output for JSON serialization.