path: location of a corpus file with whitespace-delimited tokens and ￨-delimited features within the token returns: the n

compute elapsed time

Alternate constructor. Args: model (onmt.modules.NMTModel): See :func:`__init__()`. fields (dict[str, torchtext.data.

Return set of symbol pairs in a word. word is represented as tuple of symbols (symbols being variable-length strings)

获取数字图像的路径和标签并返回对应列表 @para: img_root: 保存图像的根目录 @para:label_file_path: 保存图像标签数据的文件路径 .csv 或 .txt 分隔符为',' @return: 图像的路径列表和对应标签列表

Initialize for decoding. Repeat src objects `beam_size` times.

Update a legacy vocab/field format. Args: vocab: a list of (field name, torchtext.vocab.Vocab) pairs. This is the format form

Load in pretrained embeddings. Args: emb_file (str) : path to torch serialized embeddings

The vocab/fields need updated.

Preloading the model: updating internal datastructure It will effectively load the model if `load` is set

Recursively split segment into smaller units (by reversing BPE merges) until all units are either in-vocabulary, or cannot be split futher.

Helper function for relative positions attention.

Tranform the size of the tensors to fit for conv input.

Return word id for each subword token (annotate by joiner).

Return word id for each subword token (annotate by spacer).

Move the model to CPU and clear CUDA cache.

Move the model to GPU.

Translate content of ``src`` and get gold scores from ``tgt``. Args: src: See :func:`self.src_reader.read()`. tgt: Se

compute cross entropy

For a batch of input and its prediction, return a list of batch predict alignment src indice Tensor in size ``(batch, n_best,)``.

Padding predictions in batch and add BOS. Args: predictions (List[List[Tensor]]): `(batch, n_best,)`, for each src

Forward hidden state through bridge.

output1 - True - algorithm for matrix, False - vector; output2 - need reshape

Compute loss between 2 partial alignment matrix.

Create copy-vocab and numericalize with it. In-place adds ``"src_map"`` to ``example``. That is the copy-vocab numericalization of the tokeni

Make shard state dictionary for shards() to return iterable shards for efficient loss computation. Subclass must define this

Gather statistics in multi-processes cases Args: stat(:obj:onmt.utils.Statistics): a Statistics object to gather

Simple function to report training stats (if report_manager is set) see `onmt.utils.ReportManagerBase.report_training` for doc

Loads a vocabulary from the given path. Args: vocab_path (str): Path to utf-8 text file containing vocabulary. Each token sho

Simple function to report stats (if report_manager is set) see `onmt.utils.ReportManagerBase.report_step` for doc

To be overridden

Remove a checkpoint Args: name(str): name that indentifies the checkpoint (it may be a filepath)

Private helper for running the specific RNN forward pass. Must be overriden by all subclasses. Args: tgt (LongTe