Takes gold and predicted answer sets and first finds the optimal 1-1 alignment between them and gets maximum metric values over all t

Takes gold and predicted answer sets and first finds the optimal 1-1 alignment between them and gets maximum metric values over all t

Computes the logits for a batch of tokens. Parameters ---------- tokens : torch.Tensor The tokenized tex

Copies the entry at `src_path` to `dest_path`. The destination path can be a directory that does not exist (creating directories without a co

Copies the `state` . This method copies the data in `state` using `copy.deepcopy()`. If this is not appropriate for your constraint,

Do adaptive gradient clipping on a param group. If ``collect_param_metrics`` is ``True`` this will return the total number of gradie

Do global fixed gradient clipping on a param group. If ``collect_param_metrics`` is ``True`` this will return the total number of gr

This matches the decay/no-decay param group split without the gradient clipping. We clip the gradients separately for test.

This method is best effort. It may mark run paths as not (false negatives) or mark non-run paths as runs (false positives). We prioritize min

Load the state produced from `self._get_flat_param_state_to_save()`.

Merges attn_head_size dim and num_attn_heads dim into hidden dim

inps: a torch tensor of shape [batch, sequence] the size of sequence may vary from call to call returns: a torch tensor of s

inps: a torch tensor of shape [batch, sequence] the size of sequence may vary from call to call returns: a torch tensor of s

Naive torch training loop.

Prepares the dataset by loading and formatting. Returns ------- str The formatted dataset as a single st

Processes each batch of tokens. Parameters ---------- i : int The batch index. n_ctx : int

Override this to process (detokenize, strip, replace, etc.) individual documents. This can be used in a map over documents of a data

Override this to process (detokenize, strip, replace, etc.) individual documents. This can be used in a map over documents of a data

Returns an identifier for the question in the given document.

Returns an identifier for the question in the given document.

From catwalk. https://github.com/allenai/catwalk/blob/main/catwalk/model.py

From catwalk. https://github.com/allenai/catwalk/blob/main/catwalk/model.py