tokenizers/pr_2119/en/api/models.md

Models

BPE[[tokenizers.models.BPE]]

• vocab (Dict[str, int], optional) -- A dictionary of string keys and their ids {"am": 0,...}

• merges (List[Tuple[str, str]], optional) -- A list of pairs of tokens (Tuple[str, str]) [("a", "b"),...]

• cache_capacity (int, optional) -- The number of words that the BPE cache can contain. The cache allows to speed-up the process by keeping the result of the merge operations for a number of words.

• dropout (float, optional) -- A float between 0 and 1 that represents the BPE dropout to use.

• unk_token (str, optional) -- The unknown token to be used by the model.

• continuing_subword_prefix (str, optional) -- The prefix to attach to subword units that don't represent a beginning of word.

• end_of_word_suffix (str, optional) -- The suffix to attach to subword units that represent an end of word.

• fuse_unk (bool, optional) -- Whether to fuse any subsequent unknown tokens into a single one

• byte_fallback (bool, optional) -- Whether to use spm byte-fallback trick (defaults to False)

• ignore_merges (bool, optional) -- Whether or not to match tokens with the vocab before using merges. An implementation of the BPE (Byte-Pair Encoding) algorithm

Example:

>>> from tokenizers.models import BPE
>>> # Build an empty model (to be trained)
>>> model = BPE(unk_token="<unk>")
>>> # Load from vocabulary and merges files
>>> model = BPE.from_file("vocab.json", "merges.txt")

• vocab (str) -- The path to a vocab.json file

• merges (str) -- The path to a merges.txt fileBPEAn instance of BPE loaded from these files

Instantiate a BPE model from the given files.

This method is roughly equivalent to doing:

vocab, merges = BPE.read_file(vocab_filename, merges_filename)
bpe = BPE(vocab, merges)

If you don't need to keep the vocab, merges values lying around, this method is more optimized than manually calling read_file() to initialize a BPE

• vocab (str) -- The path to a vocab.json file

• merges (str) -- The path to a merges.txt fileA Tuple with the vocab and the mergesThe vocabulary and merges loaded into memory Read a vocab.json and a merges.txt files

This method provides a way to read and parse the content of these files, returning the relevant data structures. If you want to instantiate some BPE models from memory, this method gives you the expected input from the standard files.

Model[[tokenizers.models.Model]]

Base class for all models

The model represents the actual tokenization algorithm. This is the part that will contain and manage the learned vocabulary.

This class cannot be constructed directly. Please use one of the concrete models.

TrainerThe Trainer used to train this model Get the associated Trainer

Retrieve the Trainer associated to this Model.

• id (int) -- An ID to convert to a tokenstrThe token associated to the ID Get the token associated to an ID

• folder (str) -- The path to the target folder in which to save the various files

• prefix (str, optional) -- An optional prefix, used to prefix each file nameList[str]The list of saved files Save the current model

Save the current model in the given folder, using the given prefix for the various files that will get created. Any file with the same name that already exists in this folder will be overwritten.

• token (str) -- A token to convert to an IDintThe ID associated to the token Get the ID associated to a token

• sequence (str) -- A sequence to tokenizeA List of TokenThe generated tokens Tokenize a sequence

Unigram[[tokenizers.models.Unigram]]

• vocab (List[Tuple[str, float]], optional) -- A list of vocabulary items and their log-probability scores, e.g. [("am", -0.2442), ...]. If not provided, an empty model is created.

• unk_id (int, optional) -- The index of the unknown token in the vocabulary list.

• byte_fallback (bool, optional, defaults to False) -- Whether to use SentencePiece byte fallback for characters not in the vocabulary.

• alpha (float, optional) -- A float between 0 and 1 that represents the smoothing parameter (temperature) to use.

• nbest_size (int, optional) -- An integer greater than 0 that represents the maximum number of best paths to consider. If not set, it samples from the full lattice (i.e. all valid subword segmentations). An implementation of the Unigram algorithm

The Unigram algorithm is a subword tokenization algorithm based on unigram language models, as used in SentencePiece. It learns a vocabulary by starting with a large initial vocabulary and iteratively pruning it using the EM algorithm.

Example:

>>> from tokenizers.models import Unigram
>>> # Build an empty model (to be trained)
>>> model = Unigram()
>>> # Build from a vocabulary list
>>> vocab = [("<unk>", 0.0), ("hello", -1.0), ("world", -1.5)]
>>> model = Unigram(vocab=vocab, unk_id=0)

WordLevel[[tokenizers.models.WordLevel]]

• vocab (str, optional) -- A dictionary of string keys and their ids {"am": 0,...}

• unk_token (str, optional) -- The unknown token to be used by the model. An implementation of the WordLevel algorithm

Most simple tokenizer model based on mapping tokens to their corresponding id.

Example:

>>> from tokenizers.models import WordLevel
>>> # Build from a vocabulary dictionary
>>> vocab = {"hello": 0, "world": 1, "<unk>": 2}
>>> model = WordLevel(vocab=vocab, unk_token="<unk>")
>>> # Load from file
>>> model = WordLevel.from_file("vocab.json", unk_token="<unk>")

• vocab (str) -- The path to a vocab.json fileWordLevelAn instance of WordLevel loaded from file

Instantiate a WordLevel model from the given file

This method is roughly equivalent to doing:

vocab = WordLevel.read_file(vocab_filename)
wordlevel = WordLevel(vocab)

If you don't need to keep the vocab values lying around, this method is more optimized than manually calling read_file() to initialize a WordLevel

• vocab (str) -- The path to a vocab.json fileDict[str, int]The vocabulary as a dict Read a vocab.json

This method provides a way to read and parse the content of a vocabulary file, returning the relevant data structures. If you want to instantiate some WordLevel models from memory, this method gives you the expected input from the standard files.

WordPiece[[tokenizers.models.WordPiece]]

• vocab (Dict[str, int], optional) -- A dictionary of string keys and their ids {"am": 0,...}

• unk_token (str, optional) -- The unknown token to be used by the model.

• max_input_chars_per_word (int, optional) -- The maximum number of characters to authorize in a single word. An implementation of the WordPiece algorithm

Example:

>>> from tokenizers.models import WordPiece
>>> # Build an empty model (to be trained)
>>> model = WordPiece(unk_token="[UNK]")
>>> # Load from a vocabulary file
>>> model = WordPiece.from_file("vocab.txt")

• vocab (str) -- The path to a vocab.txt fileWordPieceAn instance of WordPiece loaded from file

Instantiate a WordPiece model from the given file

This method is roughly equivalent to doing:

vocab = WordPiece.read_file(vocab_filename)
wordpiece = WordPiece(vocab)

If you don't need to keep the vocab values lying around, this method is more optimized than manually calling read_file() to initialize a WordPiece

• vocab (str) -- The path to a vocab.txt fileDict[str, int]The vocabulary as a dict Read a vocab.txt file

This method provides a way to read and parse the content of a standard vocab.txt file as used by the WordPiece Model, returning the relevant data structures. If you want to instantiate some WordPiece models from memory, this method gives you the expected input from the standard files.

The Rust API Reference is available directly on the Docs.rs website.

The node API has not been documented yet.