Format refs and preds for sacrebleu corpus calculation. It is very particular

Create a filtering pipeline.

Copy current averaged parameters into given collection of parameters. Args: parameters: Iterable of `torch.nn.Parameter`;

Restores the original order of a new array based on the old array's order Args: newarr (List[Any]): The array to be restored

Groups elements of an iterable based on a provided function. Parameters: - arr (Iterable): The iterable to be grouped.

Whether the task has a test set

Query OpenAI API for completion. Retry with back-off until they respond

Register a tensor with a configurable learning rate and 0 weight decay

Restore the parameters stored with the `store` method. Useful to validate the model with EMA parameters without affecting the

x: (batch, dim)

num_heads_kv: can be used to toggle MQA / GQA. If None, use num_heads. return_residual: whether to return the input x along with the

Fast path that combine 3 steps: apply rotary to Q and K, update kv cache, and apply attention. q: (batch_size, seqlen_q, nheads, head

Fast path that combine 3 steps: apply rotary to Q and K, update kv cache, and apply attention. q: (batch_size, seqlen_q, nheads, head

Fast path that combine 3 steps: apply rotary to Q and K, update kv cache, and apply attention. q: (batch_size, seqlen_q, nheads, head

Returns the state tensors for the given layer. Adds support for nested states.

:param inps: torch.Tensor A torch tensor of shape [batch, (sequence_ctx + sequence_cont)] or of shape [batch, sequenc

Reorders the elements in the array based on the sorting function. Parameters: - arr (Union[List, Tuple[Tuple[int, Any], ...]

kv: (batch_size, seqlen, 2, nheads, head_dim) or (batch_size, 1, 2, nheads, head_dim)

Write kv to inference_params, then do attention

Write kv to inference_params, then do attention

Write kv to inference_params, then do attention

Ignore capitalization and determiners.

This function is copy of `random_spans_helper <https://github.com/google-research/text-to-text-transfer-transformer/blob/84f8bcc14b5f2c03de51bd3587609

Uses RequestFactory to construct Requests and returns an iterable of Requests which will be sent to the LM. :param doc: T

Sentinel ids creation given the indices that should be masked. The start indices of each mask are replaced by the sentinel ids in inc

SA: 051024 Only use encoder region for prefill in JRT.

Downloads and returns the task dataset. Override this method to download the dataset from a custom API. :param data_dir: str

Extended JSON dump that can handle pydantic models.

Fetches run data from a W&B project into a pandas DataFrame. Parameters: - project_name (str): The name of the W&B project.

Returns a fewshot context string that is made up of a prepended description (if provided), the `num_fewshot` number of examples, and an append

Puts sentinel mask on `input_ids` and fuse consecutive mask tokens into a single mask token by deleting. This will reduce the sequenc