Transform a batched tensor of document lengths into a 1D tensor of cumulative document lengths for the whole batch.

load dataset and build dataloader :param data_path_or_name: str, dataset name in hf-hub or local file path :param prompt_col_name: str, see `

Get the size of the file at the given path in bytes. Raises an error if the path does not correspond to a file.

Takes a predicted answer and a gold answer (that are both either a string or a list of strings), and returns exact match and the DROP

Takes a predicted answer and a gold answer (that are both either a string or a list of strings), and returns exact match and the DROP

Separate parameters into weight decay and non weight decay groups.

Process results from OpenAI API response. :param response: dict OpenAI API Response :param ctxlen: int Length of context (so

Process results from OpenAI API response. :param response: dict OpenAI API Response :param ctxlen: int Length of context (so

Returns the split with the given name.

Returns the split with the given name.

Function to group consecutive values in an array.

Sample `Gumbel(phi)`. `phi` should have shape `(batch_size, num_classes)`.

Returns ``True`` if a split with the given name exists. ``False`` otherwise.

Returns ``True`` if a split with the given name exists. ``False`` otherwise.

Format a number using K for thousands, M for millions, B for billions, T for trillions. Args: num: Number to format decimal_

Format a number using K for thousands, M for millions, B for billions, T for trillions. Args: num: Number to format decimal_