flair.datasets.sequence_labeling.NER_MULTI_WIKIANN#

class flair.datasets.sequence_labeling.NER_MULTI_WIKIANN(languages='en', base_path=None, in_memory=False, **corpusargs)View on GitHub #

__init__(languages='en', base_path=None, in_memory=False, **corpusargs)View on GitHub #

Initialize the WkiAnn corpus for cross-lingual NER consisting of datasets from 282 languages that exist in Wikipedia.

See https://elisa-ie.github.io/wikiann/ for details and for the languages and their respective abbreveations, i.e. “en” for english. (license: https://opendatacommons.org/licenses/by/)

Parameters:

languages (Union[str, list[str]]) – Should be an abbreviation of a language (“en”, “de”,..) or a list of abbreviations. The datasets of all passed languages will be saved in one MultiCorpus. (Note that, even though listed on https://elisa-ie.github.io/wikiann/ some datasets are empty. This includes “aa”, “cho”, “ho”, “hz”, “ii”, “jam”, “kj”, “kr”, “mus”, “olo” and “tcy”.)
base_path (Union[str, Path], optional) – Default is None, meaning that corpus gets auto-downloaded and loaded. You can override this to point to a different folder but typically this should not be necessary. The data is in bio-format. It will by default (with the string “ner” as value) be transformed into the bioes format. If you dont want that set it to None.
in_memory (bool, optional) – Specify that the dataset should be loaded in memory, which speeds up the training process but takes increases the RAM usage significantly.

Methods

`__init__`([languages, base_path, in_memory])	Initialize the WkiAnn corpus for cross-lingual NER consisting of datasets from 282 languages that exist in Wikipedia.
`add_label_noise`(label_type, labels[, ...])	Generates uniform label noise distribution in the chosen dataset split.
`downsample`([percentage, downsample_train, ...])	Randomly downsample the corpus to the given percentage (by removing data points).
`filter_empty_sentences`()	A method that filters all sentences consisting of 0 tokens.
`filter_long_sentences`(max_charlength)	A method that filters all sentences for which the plain text is longer than a specified number of characters.
`get_all_sentences`()	Returns all sentences (spanning all three splits) in the `Corpus`.
`get_label_distribution`()	Counts occurrences of each label in the corpus and returns them as a dictionary object.
`make_label_dictionary`(label_type[, ...])	Creates a dictionary of all labels assigned to the sentences in the corpus.
`make_tag_dictionary`(tag_type)	Create a tag dictionary of a given label type.
`make_vocab_dictionary`([max_tokens, min_freq])	Creates a `Dictionary` of all tokens contained in the corpus.
`obtain_statistics`([label_type, pretty_print])	Print statistics about the corpus, including the length of the sentences and the labels in the corpus.

Attributes

`dev`	The dev split as a `torch.utils.data.Dataset` object.
`test`	The test split as a `torch.utils.data.Dataset` object.
`train`	The training split as a `torch.utils.data.Dataset` object.