Perform pooling by stride slicing the tensor along the given axis.

Perform pooling by stride slicing the tensor along the given axis.

Serializes this instance to a Python dictionary.

Scikit / Keras interface to transformers' pipelines. This method will forward to __call__().

Check if the model class is in supported by the pipeline. Args: supported_models (:obj:`List[str]` or :obj:`dict`):

Context Manager allowing tensor allocation on the user-specified device in framework agnostic way. Returns: Context mana

Invert an attention mask (e.g., switches 0. and 1.). Args: encoder_attention_mask (:obj:`torch.Tensor`): An attention ma

Used to implement attention between consecutive chunks. Args: vectors: array of shape [batch_size, num_attention_heads, n_chunks,

Ensure PyTorch tensors are on the specified device. Args: inputs (keyword arguments that should be :obj:`torch.Tensor`):

Prune a Conv1D layer to keep only entries in index. A Conv1D work as a Linear layer (see e.g. BERT) but the weights are transposed. Used

Will save the model, so you can reload it using :obj:`from_pretrained()`. Will only save from the world_master process (unless in TP

splits hidden_size dim into attn_head_size and num_attn_heads

Run evaluation and returns metrics. The calling script will be responsible for providing a method to compute metrics, as they are

Run evaluation and returns metrics. The calling script will be responsible for providing a method to compute metrics, as they are

r""" Instantiates one of the base model classes of the library from a configuration. Note: Loading a model from its confi

Resolve a model identifier, and a file name, to a HF-hosted url on either S3 or Cloudfront (a Content Delivery Network, or CDN). Cloudfr

Numpify `tensors` (even if it's a nested list/tuple of tensors).

Main training entry point. Args: model_path (:obj:`str`, `optional`): Local path to the model if the mod

Creates examples for the training, dev and test sets.

Parse arguments and tokenize

Add a list of new tokens to the tokenizer class. If the new tokens are not in the vocabulary, they are added to it with indices start

From a ``pretrained_model_name_or_path``, resolve to a dictionary of parameters, to be used for instantiating a :class:`~transformers

Get token or None if not existent.

Whether or not this process is the local (e.g., on one machine if training in a distributed fashion on several machines) main process

list(map(f, x))

Helper to get number of samples in a :class:`~torch.utils.data.DataLoader` by accessing its dataset.

Helper to get number of samples in a :class:`~torch.utils.data.DataLoader` by accessing its dataset.

Helper to get number of samples in a :class:`~torch.utils.data.DataLoader` by accessing its dataset.

Just EOS

Will save the model, so you can reload it using :obj:`from_pretrained()`. Will only save from the world_master process (unless in TP

projection

Main training entry point. Args: model_path (:obj:`str`, `optional`): Local path to the model if the mod

Find the correct padding/truncation strategy with backward compatibility for old arguments (truncation_strategy and pad_to_max_length

Tie or clone module weights depending of whether we are using TorchScript or not

Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and adding special tokens

:obj:`torch.device`: The device on which the module is (assuming that all the module parameters are on the same device).

Gaussian Error Linear Unit. Original Implementation of the gelu activation function in Google Bert repo when initially created. For inform

Group together the adjacent tokens with the same entity predicted. Args: entities (:obj:`dict`): The entities predicted

Turns 1->0, 0->1, False->True, True-> False

Whether or not this process is the global main process (when training in a distributed fashion on several machines, this is only goin

Concat the `new_tensors` to `tensors` on `dim`. Works for tensors or nested list/tuples of tensors.

Lower text and remove punctuation, articles and extra whitespace.

Apply 1D pooling to a tensor of size [B x T (x H)].

Apply 1D pooling to a tensor of size [B x T (x H)].