BiaffineDependencyParserMultiLang(self,, text_field_embedder:allennlp.modules.text_field_embedders.text_field_embedder.TextFieldEmbedder, encoder:allennlp.modules.seq2seq_encoders.seq2seq_encoder.Seq2SeqEncoder, tag_representation_dim:int, arc_representation_dim:int, tag_feedforward:allennlp.modules.feedforward.FeedForward=None, arc_feedforward:allennlp.modules.feedforward.FeedForward=None, pos_tag_embedding:allennlp.modules.token_embedders.embedding.Embedding=None, use_mst_decoding_for_validation:bool=True, langs_for_early_stop:List[str]=None, dropout:float=0.0, input_dropout:float=0.0, initializer:allennlp.nn.initializers.InitializerApplicator=<allennlp.nn.initializers.InitializerApplicator object at 0x7f7237f92470>, regularizer:Union[allennlp.nn.regularizers.regularizer_applicator.RegularizerApplicator, NoneType]=None) -> None

This dependency parser implements the multi-lingual extension of the Dozat and Manning (2016) model as described in [Cross-Lingual Alignment of Contextual Word Embeddings, with Applications to Zero-shot Dependency Parsing (Schuster et al., 2019)] ( Also, please refer to the [alignment computation code] (

All parameters are shared across all languages except for the text_field_embedder. For aligned ELMo embeddings, use the elmo_token_embedder_multilang with the pre-computed alignments to the mutual embedding space. Also, the universal_dependencies_multilang dataset reader supports loading of multiple sources and storing the language identifier in the metadata.


  • vocab : Vocabulary, required A Vocabulary, required in order to compute sizes for input/output projections.
  • text_field_embedder : TextFieldEmbedder, required Used to embed the tokens TextField we get as input to the model.
  • encoder : Seq2SeqEncoder The encoder (with its own internal stacking) that we will use to generate representations of tokens.
  • tag_representation_dim : int, required. The dimension of the MLPs used for dependency tag prediction.
  • arc_representation_dim : int, required. The dimension of the MLPs used for head arc prediction.
  • tag_feedforward : FeedForward, optional, (default = None). The feedforward network used to produce tag representations. By default, a 1 layer feedforward network with an elu activation is used.
  • arc_feedforward : FeedForward, optional, (default = None). The feedforward network used to produce arc representations. By default, a 1 layer feedforward network with an elu activation is used.
  • pos_tag_embedding : Embedding, optional. Used to embed the pos_tags SequenceLabelField we get as input to the model.
  • use_mst_decoding_for_validation : bool, optional (default = True). Whether to use Edmond's algorithm to find the optimal minimum spanning tree during validation. If false, decoding is greedy.
  • langs_for_early_stop : List[str], optional, (default = []) Which languages to include in the averaged metrics (that could be used for early stopping).
  • dropout : float, optional, (default = 0.0) The variational dropout applied to the output of the encoder and MLP layers.
  • input_dropout : float, optional, (default = 0.0) The dropout applied to the embedded text input.
  • initializer : InitializerApplicator, optional (default=InitializerApplicator()) Used to initialize the model parameters.
  • regularizer : RegularizerApplicator, optional (default=None) If provided, will be used to calculate the regularization penalty during training.


BiaffineDependencyParserMultiLang.forward(self, words:Dict[str, Dict[str, torch.Tensor]], pos_tags:torch.LongTensor, metadata:List[Dict[str, Any]], head_tags:torch.LongTensor=None, head_indices:torch.LongTensor=None) -> Dict[str, torch.Tensor]

Embedding each language by the corresponding parameters for TextFieldEmbedder. Batches should contain only samples from a single language. Metadata should have a lang key.


BiaffineDependencyParserMultiLang.get_metrics(self, reset:bool=False) -> Dict[str, float]

Returns a dictionary of metrics. This method will be called by in order to compute and use model metrics for early stopping and model serialization. We return an empty dictionary here rather than raising as it is not required to implement metrics for a new model. A boolean reset parameter is passed, as frequently a metric accumulator will have some state which should be reset between epochs. This is also compatible with Metrics should be populated during the call to forward, with the handling the accumulation of the metric until this method is called.