Overview

Data formats

Details on spaCy's input and output data formats

This section documents input and output formats of data used by spaCy, including the training config, training data and lexical vocabulary data. For an overview of label schemes used by the models, see the models directory. Each trained pipeline documents the label schemes used in its components, depending on the data it was trained on.

Training config v3.0

Config files define the training process and pipeline and can be passed to spacy train. They use Thinc’s configuration system under the hood. For details on how to use training configs, see the usage documentation. To get started with the recommended settings for your use case, check out the quickstart widget or run the init config command.

explosion/spaCy/master/spacy/default_config.cfg

nlp section

Defines the nlp object, its tokenizer and processing pipeline component names.

NameDescription
langPipeline language ISO code. Defaults to null. str
pipelineNames of pipeline components in order. Should correspond to sections in the [components] block, e.g. [components.ner]. See docs on defining components. Defaults to []. List[str]
disabledNames of pipeline components that are loaded but disabled by default and not run as part of the pipeline. Should correspond to components listed in pipeline. After a pipeline is loaded, disabled components can be enabled using Language.enable_pipe. List[str]
before_creationOptional callback to modify Language subclass before it’s initialized. Defaults to null. Optional[Callable[[Type[Language]], Type[Language]]]
after_creationOptional callback to modify nlp object right after it’s initialized. Defaults to null. Optional[Callable[[Language],Language]]
after_pipeline_creationOptional callback to modify nlp object after the pipeline components have been added. Defaults to null. Optional[Callable[[Language],Language]]
tokenizerThe tokenizer to use. Defaults to Tokenizer. Callable[[str],Doc]
batch_sizeDefault batch size for Language.pipe and Language.evaluate. int

components section

This section includes definitions of the pipeline components and their models, if available. Components in this section can be referenced in the pipeline of the [nlp] block. Component blocks need to specify either a factory (named function to use to create component) or a source (name of path of trained pipeline to copy components from). See the docs on defining pipeline components for details.

paths, system variables

These sections define variables that can be referenced across the other sections as variables. For example ${paths.train} uses the value of train defined in the block [paths]. If your config includes custom registered functions that need paths, you can define them here. All config values can also be overwritten on the CLI when you run spacy train, which is especially relevant for data paths that you don’t want to hard-code in your config file.

corpora section

This section defines a dictionary mapping of string keys to functions. Each function takes an nlp object and yields Example objects. By default, the two keys train and dev are specified and each refer to a Corpus. When pretraining, an additional pretrain section is added that defaults to a JsonlCorpus. You can also register custom functions that return a callable.

NameDescription
trainTraining data corpus, typically used in [training] block. Callable[[Language], Iterator[Example]]
devDevelopment data corpus, typically used in [training] block. Callable[[Language], Iterator[Example]]
pretrainRaw text for pretraining, typically used in [pretraining] block (if available). Callable[[Language], Iterator[Example]]
Any custom or alternative corpora. Callable[[Language], Iterator[Example]]

Alternatively, the [corpora] block can refer to one function that returns a dictionary keyed by the corpus names. This can be useful if you want to load a single corpus once and then divide it up into train and dev partitions.

NameDescription
corporaA dictionary keyed by string names, mapped to corpus functions that receive the current nlp object and return an iterator of Example objects. Dict[str, Callable[[Language], Iterator[Example]]]

training section

This section defines settings and controls for the training and evaluation process that are used when you run spacy train.

NameDescription
accumulate_gradientWhether to divide the batch up into substeps. Defaults to 1. int
batcherCallable that takes an iterator of Doc objects and yields batches of Docs. Defaults to batch_by_words. Callable[[Iterator[Doc], Iterator[List[Doc]]]]
before_to_diskOptional callback to modify nlp object right before it is saved to disk during and after training. Can be used to remove or reset config values or disable components. Defaults to null. Optional[Callable[[Language],Language]]
before_update v3.5Optional callback that is invoked at the start of each training step with the nlp object and a Dict containing the following entries: step, epoch. Can be used to make deferred changes to components. Defaults to null. Optional[Callable[[Language, Dict[str, Any]], None]]
dev_corpusDot notation of the config location defining the dev corpus. Defaults to corpora.dev. str
dropoutThe dropout rate. Defaults to 0.1. float
eval_frequencyHow often to evaluate during training (steps). Defaults to 200. int
frozen_componentsPipeline component names that are “frozen” and shouldn’t be initialized or updated during training. See here for details. Defaults to []. List[str]
annotating_components v3.1Pipeline component names that should set annotations on the predicted docs during training. See here for details. Defaults to []. List[str]
gpu_allocatorLibrary for cupy to route GPU memory allocation to. Can be "pytorch" or "tensorflow". Defaults to variable ${system.gpu_allocator}. str
loggerCallable that takes the nlp and stdout and stderr IO objects, sets up the logger, and returns two new callables to log a training step and to finalize the logger. Defaults to ConsoleLogger. Callable[[Language, IO, IO], [Tuple[Callable[[Dict[str, Any]], None], Callable[[], None]]]]
max_epochsMaximum number of epochs to train for. 0 means an unlimited number of epochs. -1 means that the train corpus should be streamed rather than loaded into memory with no shuffling within the training loop. Defaults to 0. int
max_stepsMaximum number of update steps to train for. 0 means an unlimited number of steps. Defaults to 20000. int
optimizerThe optimizer. The learning rate schedule and other settings can be configured as part of the optimizer. Defaults to Adam. Optimizer
patienceHow many steps to continue without improvement in evaluation score. 0 disables early stopping. Defaults to 1600. int
score_weightsScore names shown in metrics mapped to their weight towards the final weighted score. See here for details. Defaults to {}. Dict[str, float]
seedThe random seed. Defaults to variable ${system.seed}. int
train_corpusDot notation of the config location defining the train corpus. Defaults to corpora.train. str

pretraining sectionoptional

This section is optional and defines settings and controls for language model pretraining. It’s used when you run spacy pretrain.

NameDescription
max_epochsMaximum number of epochs. Defaults to 1000. int
dropoutThe dropout rate. Defaults to 0.2. float
n_save_everySaving frequency. Defaults to null. Optional[int]
objectiveThe pretraining objective. Defaults to {"type": "characters", "n_characters": 4}. Dict[str, Any]
optimizerThe optimizer. The learning rate schedule and other settings can be configured as part of the optimizer. Defaults to Adam. Optimizer
corpusDot notation of the config location defining the corpus with raw text. Defaults to corpora.pretrain. str
batcherCallable that takes an iterator of Doc objects and yields batches of Docs. Defaults to batch_by_words. Callable[[Iterator[Doc], Iterator[List[Doc]]]]
componentComponent name to identify the layer with the model to pretrain. Defaults to "tok2vec". str
layerThe specific layer of the model to pretrain. If empty, the whole model will be used. str

initialize section

This config block lets you define resources for initializing the pipeline. It’s used by Language.initialize and typically called right before training (but not at runtime). The section allows you to specify local file paths or custom functions to load data resources from, without requiring them at runtime when you load the trained pipeline back in. Also see the usage guides on the config lifecycle and custom initialization.

NameDescription
after_initOptional callback to modify the nlp object after initialization. Optional[Callable[[Language],Language]]
before_initOptional callback to modify the nlp object before initialization. Optional[Callable[[Language],Language]]
componentsAdditional arguments passed to the initialize method of a pipeline component, keyed by component name. If type annotations are available on the method, the config will be validated against them. The initialize methods will always receive the get_examples callback and the current nlp object. Dict[str, Dict[str, Any]]
init_tok2vecOptional path to pretrained tok2vec weights created with spacy pretrain. Defaults to variable ${paths.init_tok2vec}. Ignored when actually running pretraining, as you’re creating the file to be used later. Optional[str]
lookupsAdditional lexeme and vocab data from spacy-lookups-data. Defaults to null. Optional[Lookups]
tokenizerAdditional arguments passed to the initialize method of the specified tokenizer. Can be used for languages like Chinese that depend on dictionaries or trained models for tokenization. If type annotations are available on the method, the config will be validated against them. The initialize method will always receive the get_examples callback and the current nlp object. Dict[str, Any]
vectorsName or path of pipeline containing pretrained word vectors to use, e.g. created with init vectors. Defaults to null. Optional[str]
vocab_dataPath to JSONL-formatted vocabulary file to initialize vocabulary. Optional[str]

Training data

Binary training format v3.0

The main data format used in spaCy v3.0 is a binary format created by serializing a DocBin, which represents a collection of Doc objects. This means that you can train spaCy pipelines using the same format it outputs: annotated Doc objects. The binary format is extremely efficient in storage, especially when packing multiple documents together.

Typically, the extension for these binary files is .spacy, and they are used as input format for specifying a training corpus and for spaCy’s CLI train command. The built-in convert command helps you convert spaCy’s previous JSON format to the new binary format. It also supports conversion of the .conllu format used by the Universal Dependencies corpora.

Note that while this is the format used to save training data, you do not have to understand the internal details to use it or create training data. See the section on preparing training data.

JSON training format deprecated

Example structure

Here’s an example of dependencies, part-of-speech tags and named entities, taken from the English Wall Street Journal portion of the Penn Treebank:

explosion/spaCy/v2.3.x/examples/training/training-data.json

Annotation format for creating training examples

An Example object holds the information for one training instance. It stores two Doc objects: one for holding the gold-standard reference data, and one for holding the predictions of the pipeline. Examples can be created using the Example.from_dict method with a reference Doc and a dictionary of gold-standard annotations.

NameDescription
textRaw text. str
wordsList of gold-standard tokens. List[str]
lemmasList of lemmas. List[str]
spacesList of boolean values indicating whether the corresponding tokens is followed by a space or not. List[bool]
tagsList of fine-grained POS tags. List[str]
posList of coarse-grained POS tags. List[str]
morphsList of morphological features. List[str]
sent_startsList of boolean values indicating whether each token is the first of a sentence or not. List[bool]
depsList of string values indicating the dependency relation of a token to its head. List[str]
headsList of integer values indicating the dependency head of each token, referring to the absolute index of each token in the text. List[int]
entitiesOption 1: List of BILUO tags per token of the format "{action}-{label}", or None for unannotated tokens. List[str]
entitiesOption 2: List of (start_char, end_char, label) tuples defining all entities in the text. List[Tuple[int, int, str]]
catsDictionary of label/value pairs indicating how relevant a certain text category is for the text. Dict[str, float]
linksDictionary of offset/dict pairs defining named entity links. The character offsets are linked to a dictionary of relevant knowledge base IDs. Dict[Tuple[int, int], Dict]
spansDictionary of spans_key/List[Tuple] pairs defining the spans for each spans key as (start_char, end_char, label, kb_id) tuples. Dict[str, List[Tuple[int, int, str, str]]

Examples

Lexical data for vocabulary

This data file can be provided via the vocab_data setting in the [initialize] block of the training config to pre-define the lexical data to initialize the nlp object’s vocabulary with. The file should contain one lexical entry per line. The first line defines the language and vocabulary settings. All other lines are expected to be JSON objects describing an individual lexeme. The lexical attributes will be then set as attributes on spaCy’s Lexeme object.

First line

Entry structure

Here’s an example of the 20 most frequent lexemes in the English training data:

explosion/spaCy/master/extra/example_data/vocab-data.jsonl

Pipeline meta

The pipeline meta is available as the file meta.json and exported automatically when you save an nlp object to disk. Its contents are available as nlp.meta.

NameDescription
langPipeline language ISO code. Defaults to "en". str
namePipeline name, e.g. "core_web_sm". The final package name will be {lang}_{name}. Defaults to "pipeline". str
versionPipeline version. Will be used to version a Python package created with spacy package. Defaults to "0.0.0". str
spacy_versionspaCy version range the package is compatible with. Defaults to the spaCy version used to create the pipeline, up to next minor version, which is the default compatibility for the available trained pipelines. For instance, a pipeline trained with v3.0.0 will have the version range ">=3.0.0,<3.1.0". str
parent_packageName of the spaCy package. Typically "spacy" or "spacy_nightly". Defaults to "spacy". str
requirementsPython package requirements that the pipeline depends on. Will be used for the Python package setup in spacy package. Should be a list of package names with optional version specifiers, just like you’d define them in a setup.cfg or requirements.txt. Defaults to []. List[str]
descriptionPipeline description. Also used for Python package. Defaults to "". str
authorPipeline author name. Also used for Python package. Defaults to "". str
emailPipeline author email. Also used for Python package. Defaults to "". str
urlPipeline author URL. Also used for Python package. Defaults to "". str
licensePipeline license. Also used for Python package. Defaults to "". str
sourcesData sources used to train the pipeline. Typically a list of dicts with the keys "name", "url", "author" and "license". See here for examples. Defaults to None. Optional[List[Dict[str, str]]]
vectorsInformation about the word vectors included with the pipeline. Typically a dict with the keys "width", "vectors" (number of vectors), "keys" and "name". Dict[str, Any]
pipelineNames of pipeline component names, in order. Corresponds to nlp.pipe_names. Only exists for reference and is not used to create the components. This information is defined in the config.cfg. Defaults to []. List[str]
labelsLabel schemes of the trained pipeline components, keyed by component name. Corresponds to nlp.pipe_labels. See here for examples. Defaults to {}. Dict[str, Dict[str, List[str]]]
performanceTraining accuracy, added automatically by spacy train. Dictionary of score names mapped to scores. Defaults to {}. Dict[str, Union[float, Dict[str, float]]]
speedInference speed, added automatically by spacy train. Typically a dictionary with the keys "cpu", "gpu" and "nwords" (words per second). Defaults to {}. Dict[str, Optional[Union[float, str]]]
spacy_git_version v3.0Git commit of spacy used to create pipeline. str
otherAny other custom meta information you want to add. The data is preserved in nlp.meta. Any