Version: 3.x

Components

Components make up your NLU pipeline and work sequentially to process user input into structured output. There are components for entity extraction, for intent classification, response selection, pre-processing, and more.

Language Models

The following components load pre-trained models that are needed if you want to use pre-trained word vectors in your pipeline.

MitieNLP

  • Short

    MITIE initializer

  • Outputs

    Nothing

  • Requires

    Nothing

  • Description

    Initializes MITIE structures. Every MITIE component relies on this, hence this should be put at the beginning of every pipeline that uses any MITIE components.

  • Configuration

    The MITIE library needs a language model file, that must be specified in the configuration:

    pipeline:
    - name: "MitieNLP"
    # language model to load
    model: "data/total_word_feature_extractor.dat"

    For more information where to get that file from, head over to installing MITIE.

You can also pre-train your own word vectors from a language corpus using MITIE. To do so:

  1. Get a clean language corpus (a Wikipedia dump works) as a set of text files.

  2. Build and run MITIE Wordrep Tool on your corpus. This can take several hours/days depending on your dataset and your workstation. You'll need something like 128GB of RAM for wordrep to run – yes, that's a lot: try to extend your swap.

  3. Set the path of your new total_word_feature_extractor.dat as the model parameter to the MitieNLP component in your configuration file.

    For a full example of how to train MITIE word vectors, check out 用Rasa NLU构建自己的中文NLU系统, a blogpost that goes through creating a MITIE model from a Chinese Wikipedia dump.

SpacyNLP

  • Short

    spaCy language initializer

  • Outputs

    Nothing

  • Requires

    Nothing

  • Description

    Initializes spaCy structures. Every spaCy component relies on this, hence this should be put at the beginning of every pipeline that uses any spaCy components.

  • Configuration

    You need to specify the language model to use. The name will be passed to spacy.load(name). You can find more information on the available models on the spaCy documentation.

    pipeline:
    - name: "SpacyNLP"
    # language model to load
    model: "en_core_web_md"
    # when retrieving word vectors, this will decide if the casing
    # of the word is relevant. E.g. `hello` and `Hello` will
    # retrieve the same vector, if set to `False`. For some
    # applications and models it makes sense to differentiate
    # between these two words, therefore setting this to `True`.
    case_sensitive: False

    For more information on how to download the spaCy models, head over to installing SpaCy.

    In addition to SpaCy's pretrained language models, you can also use this component to attach spaCy models that you've trained yourself.

Tokenizers

Tokenizers split text into tokens. If you want to split intents into multiple labels, e.g. for predicting multiple intents or for modeling hierarchical intent structure, use the following flags with any tokenizer:

  • intent_tokenization_flag indicates whether to tokenize intent labels or not. Set it to True, so that intent labels are tokenized.

  • intent_split_symbol sets the delimiter string to split the intent labels, default is underscore (_).

WhitespaceTokenizer

  • Short

    Tokenizer using whitespaces as a separator

  • Outputs

    tokens for user messages, responses (if present), and intents (if specified)

  • Requires

    Nothing

  • Description

    Creates a token for every whitespace separated character sequence.

    Any character not in: a-zA-Z0-9_#@& will be substituted with whitespace before splitting on whitespace if the character fulfills any of the following conditions:

    • the character follows a whitespace: " !word""word"
    • the character precedes a whitespace: "word! ""word"
    • the character is at the beginning of the string: "!word""word"
    • the character is at the end of the string: "word!""word"

    Note that:

    • "wo!rd""wo!rd"

    In addition, any character not in: a-zA-Z0-9_#@&.~:\/?[]()!$*+,;=- will be substituted with whitespace before splitting on whitespace if the character is not between numbers:

    • "twenty{one""twenty", "one" ("{"` is not between numbers)
    • "20{1""20{1" ("{"` is between numbers)

    Note that:

    • "name@example.com""name@example.com"
    • "10,000.1""10,000.1"
    • "1 - 2""1","2"
  • Configuration

    pipeline:
    - name: "WhitespaceTokenizer"
    # Flag to check whether to split intents
    "intent_tokenization_flag": False
    # Symbol on which intent should be split
    "intent_split_symbol": "_"
    # Regular expression to detect tokens
    "token_pattern": None

JiebaTokenizer

  • Short

    Tokenizer using Jieba for Chinese language

  • Outputs

    tokens for user messages, responses (if present), and intents (if specified)

  • Requires

    Nothing

  • Description

    Creates tokens using the Jieba tokenizer specifically for Chinese language. It will only work for the Chinese language.

    note

    To use JiebaTokenizer you need to install Jieba with pip3 install jieba.

  • Configuration

    User's custom dictionary files can be auto loaded by specifying the files' directory path via dictionary_path. If the dictionary_path is None (the default), then no custom dictionary will be used.

    pipeline:
    - name: "JiebaTokenizer"
    dictionary_path: "path/to/custom/dictionary/dir"
    # Flag to check whether to split intents
    "intent_tokenization_flag": False
    # Symbol on which intent should be split
    "intent_split_symbol": "_"
    # Regular expression to detect tokens
    "token_pattern": None

MitieTokenizer

  • Short

    Tokenizer using MITIE

  • Outputs

    tokens for user messages, responses (if present), and intents (if specified)

  • Description

    Creates tokens using the MITIE tokenizer.

  • Configuration

    pipeline:
    - name: "MitieTokenizer"
    # Flag to check whether to split intents
    "intent_tokenization_flag": False
    # Symbol on which intent should be split
    "intent_split_symbol": "_"
    # Regular expression to detect tokens
    "token_pattern": None

SpacyTokenizer

  • Short

    Tokenizer using spaCy

  • Outputs

    tokens for user messages, responses (if present), and intents (if specified)

  • Description

    Creates tokens using the spaCy tokenizer.

  • Configuration

    pipeline:
    - name: "SpacyTokenizer"
    # Flag to check whether to split intents
    "intent_tokenization_flag": False
    # Symbol on which intent should be split
    "intent_split_symbol": "_"
    # Regular expression to detect tokens
    "token_pattern": None

Featurizers

Text featurizers are divided into two different categories: sparse featurizers and dense featurizers. Sparse featurizers are featurizers that return feature vectors with a lot of missing values, e.g. zeros. As those feature vectors would normally take up a lot of memory, we store them as sparse features. Sparse features only store the values that are non zero and their positions in the vector. Thus, we save a lot of memory and are able to train on larger datasets.

All featurizers can return two different kind of features: sequence features and sentence features. The sequence features are a matrix of size (number-of-tokens x feature-dimension). The matrix contains a feature vector for every token in the sequence. This allows us to train sequence models. The sentence features are represented by a matrix of size (1 x feature-dimension). It contains the feature vector for the complete utterance. The sentence features can be used in any bag-of-words model. The corresponding classifier can therefore decide what kind of features to use. Note: The feature-dimension for sequence and sentence features does not have to be the same.

MitieFeaturizer

  • Short

    Creates a vector representation of user message and response (if specified) using the MITIE featurizer.

  • Outputs

    dense_features for user messages and responses

  • Type

    Dense featurizer

  • Description

    Creates features for entity extraction, intent classification, and response classification using the MITIE featurizer.

    note

    NOT used by the MitieIntentClassifier component. But can be used by any component later in the pipeline that makes use of dense_features.

  • Configuration

    The sentence vector, i.e. the vector of the complete utterance, can be calculated in two different ways, either via mean or via max pooling. You can specify the pooling method in your configuration file with the option pooling. The default pooling method is set to mean.

    pipeline:
    - name: "MitieFeaturizer"
    # Specify what pooling operation should be used to calculate the vector of
    # the complete utterance. Available options: 'mean' and 'max'.
    "pooling": "mean"

SpacyFeaturizer

  • Short

    Creates a vector representation of user message and response (if specified) using the spaCy featurizer.

  • Outputs

    dense_features for user messages and responses

  • Type

    Dense featurizer

  • Description

    Creates features for entity extraction, intent classification, and response classification using the spaCy featurizer.

  • Configuration

    The sentence vector, i.e. the vector of the complete utterance, can be calculated in two different ways, either via mean or via max pooling. You can specify the pooling method in your configuration file with the option pooling. The default pooling method is set to mean.

    pipeline:
    - name: "SpacyFeaturizer"
    # Specify what pooling operation should be used to calculate the vector of
    # the complete utterance. Available options: 'mean' and 'max'.
    "pooling": "mean"

ConveRTFeaturizer

  • Short

    Creates a vector representation of user message and response (if specified) using ConveRT model.

  • Outputs

    dense_features for user messages and responses

  • Type

    Dense featurizer

  • Description

    Creates features for entity extraction, intent classification, and response selection. It uses the default signature to compute vector representations of input text.

    note

    Since ConveRT model is trained only on an English corpus of conversations, this featurizer should only be used if your training data is in English language.

    note

    Note that this component cannot currently run on MacOS using M1 / M2 architecture. More information on this limitation is available here.

  • Configuration

    pipeline:
    - name: "ConveRTFeaturizer"
    # Remote URL/Local directory of model files(Required)
    "model_url": None
    caution

    Since the public URL of the ConveRT model was taken offline recently, it is now mandatory to set the parameter model_url to a community/self-hosted URL or path to a local directory containing model files.

    :::

LanguageModelFeaturizer

  • Short

    Creates a vector representation of user message and response (if specified) using a pre-trained language model.

  • Outputs

    dense_features for user messages and responses

  • Type

    Dense featurizer

  • Description

    Creates features for entity extraction, intent classification, and response selection. Uses a pre-trained language model to compute vector representations of input text.

    note

    Please make sure that you use a language model which is pre-trained on the same language corpus as that of your training data.

  • Configuration

    Include a Tokenizer component before this component.

    You should specify what language model to load via the parameter model_name. See the below table for the currently supported language models. The weights to be loaded can be specified by the additional parameter model_weights. If left empty, it uses the default model weights listed in the table.

    +----------------+--------------+-------------------------+
    | Language Model | Parameter | Default value for |
    | | "model_name" | "model_weights" |
    +----------------+--------------+-------------------------+
    | BERT | bert | rasa/LaBSE |
    +----------------+--------------+-------------------------+
    | GPT | gpt | openai-gpt |
    +----------------+--------------+-------------------------+
    | GPT-2 | gpt2 | gpt2 |
    +----------------+--------------+-------------------------+
    | XLNet | xlnet | xlnet-base-cased |
    +----------------+--------------+-------------------------+
    | DistilBERT | distilbert | distilbert-base-uncased |
    +----------------+--------------+-------------------------+
    | RoBERTa | roberta | roberta-base |
    +----------------+--------------+-------------------------+
    | camemBERT | camembert | camembert-base |
    +----------------+--------------+-------------------------+

    Apart from the default pretrained model weights, further models can be used from HuggingFace models provided the following conditions are met (the mentioned files can be found in the "Files and versions" section of the model website):

    • The model architecture is one of the supported language models (check that the model_type in config.json is listed in the table's column model_name)
    • The model has pretrained Tensorflow weights (check that the file tf_model.h5 exists)
    • The model uses the default tokenizer (config.json should not contain a custom tokenizer_class setting)
    note

    The LaBSE weights that are loaded as default for the bert architecture provide a multi-lingual model trained on 112 languages (see our tutorial and the original paper). We strongly encourage using this as a baseline and testing your bot end-to-end before trying to optimize this component with other weights/architectures.

The following configuration loads the language model BERT with rasa/LaBSE weights, which can be found here:

pipeline:
- name: LanguageModelFeaturizer
# Name of the language model to use
model_name: "bert"
# Pre-Trained weights to be loaded
model_weights: "rasa/LaBSE"
# An optional path to a directory from which
# to load pre-trained model weights.
# If the requested model is not found in the
# directory, it will be downloaded and
# cached in this directory for future use.
# The default value of `cache_dir` can be
# set using the environment variable
# `TRANSFORMERS_CACHE`, as per the
# Transformers library.
cache_dir: null

RegexFeaturizer

  • Short

    Creates a vector representation of user message using regular expressions.

  • Outputs

    sparse_features for user messages and tokens.pattern

  • Requires

    tokens

  • Type

    Sparse featurizer

  • Description

    Creates features for entity extraction and intent classification. During training the RegexFeaturizer creates a list of regular expressions defined in the training data format. For each regex, a feature will be set marking whether this expression was found in the user message or not. All features will later be fed into an intent classifier / entity extractor to simplify classification (assuming the classifier has learned during the training phase, that this set feature indicates a certain intent / entity). Regex features for entity extraction are currently only supported by the CRFEntityExtractor and the DIETClassifier components!

  • Configuration

    Make the featurizer case insensitive by adding the case_sensitive: False option, the default being case_sensitive: True.

    To correctly process languages such as Chinese that don't use whitespace for word separation, the user needs to add the use_word_boundaries: False option, the default being use_word_boundaries: True.

    pipeline:
    - name: "RegexFeaturizer"
    # Text will be processed with case sensitive as default
    "case_sensitive": True
    # use match word boundaries for lookup table
    "use_word_boundaries": True

    Configuring for incremental training

    To ensure that sparse_features are of fixed size during incremental training, the component should be configured to account for additional patterns that may be added to the training data in future. To do so, configure the number_additional_patterns parameter while training the base model from scratch:

    pipeline:
    - name: RegexFeaturizer
    number_additional_patterns: 10

    If not configured by the user, the component will use twice the number of patterns currently present in the training data (including lookup tables and regex patterns) as the default value for number_additional_patterns. This number is kept at a minimum of 10 in order to avoid running out of additional slots for new patterns too frequently during incremental training. Once the component runs out of additional pattern slots, the new patterns are dropped and not considered during featurization. At this point, it is advisable to retrain a new model from scratch.

CountVectorsFeaturizer

  • Short

    Creates bag-of-words representation of user messages, intents, and responses.

  • Outputs

    sparse_features for user messages, intents, and responses

  • Requires

    tokens

  • Type

    Sparse featurizer

  • Description

    Creates features for intent classification and response selection. Creates bag-of-words representation of user message, intent, and response using sklearn's CountVectorizer. All tokens which consist only of digits (e.g. 123 and 99 but not a123d) will be assigned to the same feature.

  • Configuration

    See sklearn's CountVectorizer docs for detailed description of the configuration parameters.

    This featurizer can be configured to use word or character n-grams, using the analyzer configuration parameter. By default analyzer is set to word so word token counts are used as features. If you want to use character n-grams, set analyzer to char or char_wb. The lower and upper boundaries of the n-grams can be configured via the parameters min_ngram and max_ngram. By default both of them are set to 1. By default the featurizer takes the lemma of a word instead of the word directly if it is available. The lemma of a word is currently only set by the SpacyTokenizer. You can disable this behavior by setting use_lemma to False.

    note

    Option char_wb creates character n-grams only from text inside word boundaries; n-grams at the edges of words are padded with space. This option can be used to create Subword Semantic Hashing.

    note

    For character n-grams do not forget to increase min_ngram and max_ngram parameters. Otherwise the vocabulary will contain only single letters.

    Handling Out-Of-Vocabulary (OOV) words:

    note

    Enabled only if analyzer is word.

    Since the training is performed on limited vocabulary data, it cannot be guaranteed that during prediction an algorithm will not encounter an unknown word (a word that were not seen during training). In order to teach an algorithm how to treat unknown words, some words in training data can be substituted by generic word OOV_token. In this case during prediction all unknown words will be treated as this generic word OOV_token.

    For example, one might create separate intent outofscope in the training data containing messages of different number of OOV_token s and maybe some additional general words. Then an algorithm will likely classify a message with unknown words as this intent outofscope.

    You can either set the OOV_token or a list of words OOV_words:

    • OOV_token set a keyword for unseen words; if training data contains OOV_token as words in some messages, during prediction the words that were not seen during training will be substituted with provided OOV_token; if OOV_token=None (default behavior) words that were not seen during training will be ignored during prediction time;

    • OOV_words set a list of words to be treated as OOV_token during training; if a list of words that should be treated as Out-Of-Vocabulary is known, it can be set to OOV_words instead of manually changing it in training data or using custom preprocessor.

    note

    This featurizer creates a bag-of-words representation by counting words, so the number of OOV_token in the sentence might be important.

    note

    Providing OOV_words is optional, training data can contain OOV_token input manually or by custom additional preprocessor. Unseen words will be substituted with OOV_token only if this token is present in the training data or OOV_words list is provided.

    If you want to share the vocabulary between user messages and intents, you need to set the option use_shared_vocab to True. In that case a common vocabulary set between tokens in intents and user messages is build.

    pipeline:
    - name: "CountVectorsFeaturizer"
    # Analyzer to use, either 'word', 'char', or 'char_wb'
    "analyzer": "word"
    # Set the lower and upper boundaries for the n-grams
    "min_ngram": 1
    "max_ngram": 1
    # Set the out-of-vocabulary token
    "OOV_token": "_oov_"
    # Whether to use a shared vocab
    "use_shared_vocab": False

    Configuring for incremental training

To ensure that sparse_features are of fixed size during incremental training, the component should be configured to account for additional vocabulary tokens that may be added as part of new training examples in the future. To do so, configure the additional_vocabulary_size parameter while training the base model from scratch:

pipeline:
- name: CountVectorsFeaturizer
additional_vocabulary_size:
text: 1000
response: 1000
action_text: 1000

As in the above example, you can define additional vocabulary size for each of text (user messages), response (bot responses used by ResponseSelector) and action_text (bot responses not used by ResponseSelector). If you are building a shared vocabulary (use_shared_vocab=True), you only need to define a value for the text attribute. If any of the attribute is not configured by the user, the component takes half of the current vocabulary size as the default value for the attribute's additional_vocabulary_size. This number is kept at a minimum of 1000 in order to avoid running out of additional vocabulary slots too frequently during incremental training. Once the component runs out of additional vocabulary slots, the new vocabulary tokens are dropped and not considered during featurization. At this point, it is advisable to retrain a new model from scratch.

The above configuration parameters are the ones you should configure to fit your model to your data. However, additional parameters exist that can be adapted.

More configurable parameters
+---------------------------+-------------------------+--------------------------------------------------------------+
| Parameter | Default Value | Description |
+===========================+=========================+==============================================================+
| use_shared_vocab | False | If set to 'True' a common vocabulary is used for labels |
| | | and user message. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| analyzer | word | Whether the features should be made of word n-gram or |
| | | character n-grams. Option 'char_wb' creates character |
| | | n-grams only from text inside word boundaries; |
| | | n-grams at the edges of words are padded with space. |
| | | Valid values: 'word', 'char', 'char_wb'. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| strip_accents | None | Remove accents during the pre-processing step. |
| | | Valid values: 'ascii', 'unicode', 'None'. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| stop_words | None | A list of stop words to use. |
| | | Valid values: 'english' (uses an internal list of |
| | | English stop words), a list of custom stop words, or |
| | | 'None'. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| min_df | 1 | When building the vocabulary ignore terms that have a |
| | | document frequency strictly lower than the given threshold. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| max_df | 1 | When building the vocabulary ignore terms that have a |
| | | document frequency strictly higher than the given threshold |
| | | (corpus-specific stop words). |
+---------------------------+-------------------------+--------------------------------------------------------------+
| min_ngram | 1 | The lower boundary of the range of n-values for different |
| | | word n-grams or char n-grams to be extracted. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| max_ngram | 1 | The upper boundary of the range of n-values for different |
| | | word n-grams or char n-grams to be extracted. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| max_features | None | If not 'None', build a vocabulary that only consider the top |
| | | max_features ordered by term frequency across the corpus. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| lowercase | True | Convert all characters to lowercase before tokenizing. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| OOV_token | None | Keyword for unseen words. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| OOV_words | [] | List of words to be treated as 'OOV_token' during training. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| alias | CountVectorFeaturizer | Alias name of featurizer. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| use_lemma | True | Use the lemma of words for featurization. |
+---------------------------+-------------------------+--------------------------------------------------------------+
| additional_vocabulary_size| text: 1000 | Size of additional vocabulary to account for incremental |
| | response: 1000 | training while training a model from scratch |
| | action_text: 1000 | |
+---------------------------+-------------------------+--------------------------------------------------------------+

LexicalSyntacticFeaturizer

  • Short

    Creates lexical and syntactic features for a user message to support entity extraction.

  • Outputs

    sparse_features for user messages

  • Requires

    tokens

  • Type

    Sparse featurizer

  • Description

    Creates features for entity extraction. Moves with a sliding window over every token in the user message and creates features according to the configuration (see below). As a default configuration is present, you don't need to specify a configuration.

  • Configuration

    You can configure what kind of lexical and syntactic features the featurizer should extract. The following features are available:

    ============== ==========================================================================================
    Feature Name Description
    ============== ==========================================================================================
    BOS Checks if the token is at the beginning of the sentence.
    EOS Checks if the token is at the end of the sentence.
    low Checks if the token is lower case.
    upper Checks if the token is upper case.
    title Checks if the token starts with an uppercase character and all remaining characters are
    lowercased.
    digit Checks if the token contains just digits.
    prefix5 Take the first five characters of the token.
    prefix2 Take the first two characters of the token.
    suffix5 Take the last five characters of the token.
    suffix3 Take the last three characters of the token.
    suffix2 Take the last two characters of the token.
    suffix1 Take the last character of the token.
    pos Take the Part-of-Speech tag of the token (``SpacyTokenizer`` required).
    pos2 Take the first two characters of the Part-of-Speech tag of the token
    (``SpacyTokenizer`` required).
    ============== ==========================================================================================

    As the featurizer is moving over the tokens in a user message with a sliding window, you can define features for previous tokens, the current token, and the next tokens in the sliding window. You define the features as a [before, token, after] array. If you want to define features for the token before, the current token, and the token after, your features configuration would look like this:

    pipeline:
    - name: LexicalSyntacticFeaturizer
    "features": [
    ["low", "title", "upper"],
    ["BOS", "EOS", "low", "upper", "title", "digit"],
    ["low", "title", "upper"],
    ]

    This configuration is also the default configuration.

    note

    If you want to make use of pos or pos2 you need to add SpacyTokenizer to your pipeline.

Intent Classifiers

Intent classifiers assign one of the intents defined in the domain file to incoming user messages.

MitieIntentClassifier

  • Outputs

    intent

  • Requires

    tokens for user message and MitieNLP

  • Output-Example

    {
    "intent": {"name": "greet", "confidence": 0.98343}
    }
  • Description

    This classifier uses MITIE to perform intent classification. The underlying classifier is using a multi-class linear SVM with a sparse linear kernel (see train_text_categorizer_classifier function at the MITIE trainer code).

    note

    This classifier does not rely on any featurizer as it extracts features on its own.

  • Configuration

    pipeline:
    - name: "MitieIntentClassifier"

LogisticRegressionClassifier

  • Outputs

    intent and intent_ranking

  • Requires

    Either sparse_features or dense_features need to be present.

  • Output-Example
{
"intent": {"name": "greet", "confidence": 0.780},
"intent_ranking": [
{
"confidence": 0.780,
"name": "greet"
},
{
"confidence": 0.140,
"name": "goodbye"
},
{
"confidence": 0.080,
"name": "restaurant_search"
}
]
}
  • Description

    This classifier uses scikit-learn's logistic regression implementation to perform intent classification. It's able to use only sparse features, but will also pick up any dense features that are present. In general, DIET should yield higher accuracy results, but this classifier should train faster and may be used as a lightweight benchmark. Our implementation uses the base settings from scikit-learn, with the exception of the class_weight parameter where we assume the "balanced" setting.

  • Configuration

An example configuration with all the defaults can be found below.

pipeline:
- name: LogisticRegressionClassifier
max_iter: 100
solver: lbfgs
tol: 0.0001
random_state: 42
ranking_length: 10

There configuration parameters are briefly explained below.

  • max_iter: Maximum number of iterations taken for the solvers to converge.
  • solver: Solver to be used. For very small datasets you might consider liblinear.
  • tol: Tolerance for stopping criteria of the optimizer.
  • random_state: Used to shuffle the data before training.
  • ranking_length: Number of top intents to report. Set to 0 to report all intents

More details on the parameters can be found on the scikit-learn documentation page.

SklearnIntentClassifier

  • Short

    Sklearn intent classifier

  • Outputs

    intent and intent_ranking

  • Requires

    dense_features for user messages

  • Output-Example

    {
    "intent": {"name": "greet", "confidence": 0.7800},
    "intent_ranking": [
    {
    "confidence": 0.7800,
    "name": "greet"
    },
    {
    "confidence": 0.1400,
    "name": "goodbye"
    },
    {
    "confidence": 0.0800,
    "name": "restaurant_search"
    }
    ]
    }
  • Description

    The sklearn intent classifier trains a linear SVM which gets optimized using a grid search. It also provides rankings of the labels that did not “win”. The SklearnIntentClassifier needs to be preceded by a dense featurizer in the pipeline. This dense featurizer creates the features used for the classification. For more information about the algorithm itself, take a look at the GridSearchCV documentation.

  • Configuration

    During the training of the SVM a hyperparameter search is run to find the best parameter set. In the configuration you can specify the parameters that will get tried.

    pipeline:
    - name: "SklearnIntentClassifier"
    # Specifies the list of regularization values to
    # cross-validate over for C-SVM.
    # This is used with the ``kernel`` hyperparameter in GridSearchCV.
    C: [1, 2, 5, 10, 20, 100]
    # Specifies the kernel to use with C-SVM.
    # This is used with the ``C`` hyperparameter in GridSearchCV.
    kernels: ["linear"]
    # Gamma parameter of the C-SVM.
    "gamma": [0.1]
    # We try to find a good number of cross folds to use during
    # intent training, this specifies the max number of folds.
    "max_cross_validation_folds": 5
    # Scoring function used for evaluating the hyper parameters.
    # This can be a name or a function.
    "scoring_function": "f1_weighted"

KeywordIntentClassifier

  • Short

    Simple keyword matching intent classifier, intended for small, short-term projects.

  • Outputs

    intent

  • Requires

    Nothing

  • Output-Example

    {
    "intent": {"name": "greet", "confidence": 1.0}
    }
  • Description

    This classifier works by searching a message for keywords. The matching is case sensitive by default and searches only for exact matches of the keyword-string in the user message. The keywords for an intent are the examples of that intent in the NLU training data. This means the entire example is the keyword, not the individual words in the example.

    note

    This classifier is intended only for small projects or to get started. If you have few NLU training data, you can take a look at the recommended pipelines in Tuning Your Model.

  • Configuration

    pipeline:
    - name: "KeywordIntentClassifier"
    case_sensitive: True

DIETClassifier

  • Short

    Dual Intent Entity Transformer (DIET) used for intent classification and entity extraction

  • Outputs

    entities, intent and intent_ranking

  • Requires

    dense_features and/or sparse_features for user message and optionally the intent

  • Output-Example

    {
    "intent": {"name": "greet", "confidence": 0.7800},
    "intent_ranking": [
    {
    "confidence": 0.7800,
    "name": "greet"
    },
    {
    "confidence": 0.1400,
    "name": "goodbye"
    },
    {
    "confidence": 0.0800,
    "name": "restaurant_search"
    }
    ],
    "entities": [{
    "end": 53,
    "entity": "time",
    "start": 48,
    "value": "2017-04-10T00:00:00.000+02:00",
    "confidence": 1.0,
    "extractor": "DIETClassifier"
    }]
    }
  • Description

    DIET (Dual Intent and Entity Transformer) is a multi-task architecture for intent classification and entity recognition. The architecture is based on a transformer which is shared for both tasks. A sequence of entity labels is predicted through a Conditional Random Field (CRF) tagging layer on top of the transformer output sequence corresponding to the input sequence of tokens. For the intent labels the transformer output for the complete utterance and intent labels are embedded into a single semantic vector space. We use the dot-product loss to maximize the similarity with the target label and minimize similarities with negative samples.

    DIET does not provide pre-trained word embeddings or pre-trained language models but it is able to use these features if they are added to the pipeline. If you want to learn more about the model, check out the Algorithm Whiteboard series on YouTube, where we explain the model architecture in detail.

    note

    If during prediction time a message contains only words unseen during training and no Out-Of-Vocabulary preprocessor was used, an empty intent None is predicted with confidence 0.0. This might happen if you only use the CountVectorsFeaturizer with a word analyzer as featurizer. If you use the char_wb analyzer, you should always get an intent with a confidence value > 0.0.

  • Configuration

    If you want to use the DIETClassifier just for intent classification, set entity_recognition to False. If you want to do only entity recognition, set intent_classification to False. By default DIETClassifier does both, i.e. entity_recognition and intent_classification are set to True.

    You can define a number of hyperparameters to adapt the model. If you want to adapt your model, start by modifying the following parameters:

    • epochs: This parameter sets the number of times the algorithm will see the training data (default: 300). One epoch is equals to one forward pass and one backward pass of all the training examples. Sometimes the model needs more epochs to properly learn. Sometimes more epochs don't influence the performance. The lower the number of epochs the faster the model is trained.

    • hidden_layers_sizes: This parameter allows you to define the number of feed forward layers and their output dimensions for user messages and intents (default: text: [], label: []). Every entry in the list corresponds to a feed forward layer. For example, if you set text: [256, 128], we will add two feed forward layers in front of the transformer. The vectors of the input tokens (coming from the user message) will be passed on to those layers. The first layer will have an output dimension of 256 and the second layer will have an output dimension of 128. If an empty list is used (default behavior), no feed forward layer will be added. Make sure to use only positive integer values. Usually, numbers of power of two are used. Also, it is usual practice to have decreasing values in the list: next value is smaller or equal to the value before.

    • embedding_dimension: This parameter defines the output dimension of the embedding layers used inside the model (default: 20). We are using multiple embeddings layers inside the model architecture. For example, the vector of the complete utterance and the intent is passed on to an embedding layer before they are compared and the loss is calculated.

    • number_of_transformer_layers: This parameter sets the number of transformer layers to use (default: 2). The number of transformer layers corresponds to the transformer blocks to use for the model.

    • transformer_size: This parameter sets the number of units in the transformer (default: 256). The vectors coming out of the transformers will have the given transformer_size. The transformer_size should be a multiple of the number_of_attention_heads parameter, the training exits with an error otherwise.

    • connection_density: This parameter defines the fraction of kernel weights that are set to non zero values for all feed forward layers in the model (default: 0.2). The value should be between 0 and 1. If you set connection_density to 1, no kernel weights will be set to 0, the layer acts as a standard feed forward layer. You should not set connection_density to 0 as this would result in all kernel weights being 0, i.e. the model is not able to learn.

    • constrain_similarities: This parameter when set to True applies a sigmoid cross entropy loss over all similarity terms. This helps in keeping similarities between input and negative labels to smaller values. This should help in better generalization of the model to real world test sets.

    • model_confidence: This parameter allows the user to configure how confidences are computed during inference. It can take only one value as input which is softmax1. In softmax, confidences are in the range [0, 1]. The computed similarities are normalized with the softmax activation function.

The above configuration parameters are the ones you should configure to fit your model to your data. However, additional parameters exist that can be adapted.

More configurable parameters
+---------------------------------+------------------+--------------------------------------------------------------+
| Parameter | Default Value | Description |
+=================================+==================+==============================================================+
| hidden_layers_sizes | text: [] | Hidden layer sizes for layers before the embedding layers |
| | label: [] | for user messages and labels. The number of hidden layers is |
| | | equal to the length of the corresponding list. |
+---------------------------------+------------------+--------------------------------------------------------------+
| share_hidden_layers | False | Whether to share the hidden layer weights between user |
| | | messages and labels. |
+---------------------------------+------------------+--------------------------------------------------------------+
| transformer_size | 256 | Number of units in transformer. |
+---------------------------------+------------------+--------------------------------------------------------------+
| number_of_transformer_layers | 2 | Number of transformer layers. |
+---------------------------------+------------------+--------------------------------------------------------------+
| number_of_attention_heads | 4 | Number of attention heads in transformer. |
+---------------------------------+------------------+--------------------------------------------------------------+
| use_key_relative_attention | False | If 'True' use key relative embeddings in attention. |
+---------------------------------+------------------+--------------------------------------------------------------+
| use_value_relative_attention | False | If 'True' use value relative embeddings in attention. |
+---------------------------------+------------------+--------------------------------------------------------------+
| max_relative_position | None | Maximum position for relative embeddings. |
+---------------------------------+------------------+--------------------------------------------------------------+
| unidirectional_encoder | False | Use a unidirectional or bidirectional encoder. |
+---------------------------------+------------------+--------------------------------------------------------------+
| batch_size | [64, 256] | Initial and final value for batch sizes. |
| | | Batch size will be linearly increased for each epoch. |
| | | If constant `batch_size` is required, pass an int, e.g. `8`. |
+---------------------------------+------------------+--------------------------------------------------------------+
| batch_strategy | "balanced" | Strategy used when creating batches. |
| | | Can be either 'sequence' or 'balanced'. |
+---------------------------------+------------------+--------------------------------------------------------------+
| epochs | 300 | Number of epochs to train. |
+---------------------------------+------------------+--------------------------------------------------------------+
| random_seed | None | Set random seed to any 'int' to get reproducible results. |
+---------------------------------+------------------+--------------------------------------------------------------+
| learning_rate | 0.001 | Initial learning rate for the optimizer. |
+---------------------------------+------------------+--------------------------------------------------------------+
| embedding_dimension | 20 | Dimension size of embedding vectors. |
+---------------------------------+------------------+--------------------------------------------------------------+
| dense_dimension | text: 128 | Dense dimension for sparse features to use. |
| | label: 20 | |
+---------------------------------+------------------+--------------------------------------------------------------+
| concat_dimension | text: 128 | Concat dimension for sequence and sentence features. |
| | label: 20 | |
+---------------------------------+------------------+--------------------------------------------------------------+
| number_of_negative_examples | 20 | The number of incorrect labels. The algorithm will minimize |
| | | their similarity to the user input during training. |
+---------------------------------+------------------+--------------------------------------------------------------+
| similarity_type | "auto" | Type of similarity measure to use, either 'auto' or 'cosine' |
| | | or 'inner'. |
+---------------------------------+------------------+--------------------------------------------------------------+
| loss_type | "cross_entropy" | The type of the loss function, either 'cross_entropy' |
| | | or 'margin'. If type 'margin' is specified, |
| | | "model_confidence=cosine" will be used which is deprecated |
| | | as of 2.3.4. See footnote (1). |
+---------------------------------+------------------+--------------------------------------------------------------+
| ranking_length | 10 | Number of top intents to report. Set to 0 to report all |
| | | intents. |
+---------------------------------+------------------+--------------------------------------------------------------+
| renormalize_confidences | False | Normalize the reported top intents. Applicable only with loss|
| | | type 'cross_entropy' and 'softmax' confidences. |
+---------------------------------+------------------+--------------------------------------------------------------+
| maximum_positive_similarity | 0.8 | Indicates how similar the algorithm should try to make |
| | | embedding vectors for correct labels. |
| | | Should be 0.0 < ... < 1.0 for 'cosine' similarity type. |
+---------------------------------+------------------+--------------------------------------------------------------+
| maximum_negative_similarity | -0.4 | Maximum negative similarity for incorrect labels. |
| | | Should be -1.0 < ... < 1.0 for 'cosine' similarity type. |
+---------------------------------+------------------+--------------------------------------------------------------+
| use_maximum_negative_similarity | True | If 'True' the algorithm only minimizes maximum similarity |
| | | over incorrect intent labels, used only if 'loss_type' is |
| | | set to 'margin'. |
+---------------------------------+------------------+--------------------------------------------------------------+
| scale_loss | False | Scale loss inverse proportionally to confidence of correct |
| | | prediction. |
+---------------------------------+------------------+--------------------------------------------------------------+
| regularization_constant | 0.002 | The scale of regularization. |
+---------------------------------+------------------+--------------------------------------------------------------+
| negative_margin_scale | 0.8 | The scale of how important it is to minimize the maximum |
| | | similarity between embeddings of different labels. |
+---------------------------------+------------------+--------------------------------------------------------------+
| connection_density | 0.2 | Connection density of the weights in dense layers. |
| | | Value should be between 0 and 1. |
+---------------------------------+------------------+--------------------------------------------------------------+
| drop_rate | 0.2 | Dropout rate for encoder. Value should be between 0 and 1. |
| | | The higher the value the higher the regularization effect. |
+---------------------------------+------------------+--------------------------------------------------------------+
| drop_rate_attention | 0.0 | Dropout rate for attention. Value should be between 0 and 1. |
| | | The higher the value the higher the regularization effect. |
+---------------------------------+------------------+--------------------------------------------------------------+
| use_sparse_input_dropout | True | If 'True' apply dropout to sparse input tensors. |
+---------------------------------+------------------+--------------------------------------------------------------+
| use_dense_input_dropout | True | If 'True' apply dropout to dense input tensors. |
+---------------------------------+------------------+--------------------------------------------------------------+
| evaluate_every_number_of_epochs | 20 | How often to calculate validation accuracy. |
| | | Set to '-1' to evaluate just once at the end of training. |
+---------------------------------+------------------+--------------------------------------------------------------+
| evaluate_on_number_of_examples | 0 | How many examples to use for hold out validation set. |
| | | Large values may hurt performance, e.g. model accuracy. |
+---------------------------------+------------------+--------------------------------------------------------------+
| intent_classification | True | If 'True' intent classification is trained and intents are |
| | | predicted. |
+---------------------------------+------------------+--------------------------------------------------------------+
| entity_recognition | True | If 'True' entity recognition is trained and entities are |
| | | extracted. |
+---------------------------------+------------------+--------------------------------------------------------------+
| use_masked_language_model | False | If 'True' random tokens of the input message will be masked |
| | | and the model has to predict those tokens. It acts like a |
| | | regularizer and should help to learn a better contextual |
| | | representation of the input. |
+---------------------------------+------------------+--------------------------------------------------------------+
| tensorboard_log_directory | None | If you want to use tensorboard to visualize training |
| | | metrics, set this option to a valid output directory. You |
| | | can view the training metrics after training in tensorboard |
| | | via 'tensorboard --logdir <path-to-given-directory>'. |
+---------------------------------+------------------+--------------------------------------------------------------+
| tensorboard_log_level | "epoch" | Define when training metrics for tensorboard should be |
| | | logged. Either after every epoch ('epoch') or for every |
| | | training step ('batch'). |
+---------------------------------+------------------+--------------------------------------------------------------+
| featurizers | [] | List of featurizer names (alias names). Only features |
| | | coming from the listed names are used. If list is empty |
| | | all available features are used. |
+---------------------------------+------------------+--------------------------------------------------------------+
| checkpoint_model | False | Save the best performing model during training. Models are |
| | | stored to the location specified by `--out`. Only the one |
| | | best model will be saved. |
| | | Requires `evaluate_on_number_of_examples > 0` and |
| | | `evaluate_every_number_of_epochs > 0` |
+---------------------------------+------------------+--------------------------------------------------------------+
| split_entities_by_comma | True | Splits a list of extracted entities by comma to treat each |
| | | one of them as a single entity. Can either be `True`/`False` |
| | | globally, or set per entity type, such as: |
| | | ``` |
| | | ... |
| | | - name: DIETClassifier |
| | | split_entities_by_comma: |
| | | address: True |
| | | ... |
| | | ... |
| | | ``` |
+---------------------------------+------------------+--------------------------------------------------------------+
| constrain_similarities | False | If `True`, applies sigmoid on all similarity terms and adds |
| | | it to the loss function to ensure that similarity values are |
| | | approximately bounded. Used only if `loss_type=cross_entropy`|
+---------------------------------+------------------+--------------------------------------------------------------+
| model_confidence | "softmax" | Affects how model's confidence for each intent |
| | | is computed. Currently, only one value is supported: |
| | | 1. `softmax` - Similarities between input and intent |
| | | embeddings are post-processed with a softmax function, |
| | | as a result of which confidence for all intents sum up to 1. |
| | | This parameter does not affect the confidence for entity |
| | | prediction. |
+---------------------------------+------------------+--------------------------------------------------------------+
note

Parameter maximum_negative_similarity is set to a negative value to mimic the original starspace algorithm in the case maximum_negative_similarity = maximum_positive_similarity and use_maximum_negative_similarity = False. See starspace paper for details.

FallbackClassifier

  • Short

    Classifies a message with the intent nlu_fallback if the NLU intent classification scores are ambiguous. The confidence is set to be the same as the fallback threshold.

  • Outputs

    entities, intent and intent_ranking

  • Requires

    intent and intent_ranking output from a previous intent classifier

  • Output-Example

    {
    "intent": {"name": "nlu_fallback", "confidence": 0.7183846840434321},
    "intent_ranking": [
    {
    "confidence": 0.7183846840434321,
    "name": "nlu_fallback"
    },
    {
    "confidence": 0.28161531595656784,
    "name": "restaurant_search"
    }
    ],
    "entities": [{
    "end": 53,
    "entity": "time",
    "start": 48,
    "value": "2017-04-10T00:00:00.000+02:00",
    "confidence": 1.0,
    "extractor": "DIETClassifier"
    }]
    }
  • Description

    The FallbackClassifier classifies a user message with the intent nlu_fallback in case the previous intent classifier wasn't able to classify an intent with a confidence greater or equal than the threshold of the FallbackClassifier. It can also predict the fallback intent in the case when the confidence scores of the two top ranked intents are closer than the the ambiguity_threshold.

    You can use the FallbackClassifier to implement a Fallback Action which handles message with uncertain NLU predictions.

    rules:
    - rule: Ask the user to rephrase in case of low NLU confidence
    steps:
    - intent: nlu_fallback
    - action: utter_please_rephrase
  • Configuration

    The FallbackClassifier will only add its prediction for the nlu_fallback intent in case no other intent was predicted with a confidence greater or equal than threshold.

    • threshold: This parameter sets the threshold for predicting the nlu_fallback intent. If no intent predicted by a previous intent classifier has a confidence level greater or equal than threshold the FallbackClassifier will add a prediction of the nlu_fallback intent with a confidence 1.0.
    • ambiguity_threshold: If you configure an ambiguity_threshold, the FallbackClassifier will also predict the nlu_fallback intent in case the difference of the confidence scores for the two highest ranked intents is smaller than the ambiguity_threshold.

Entity Extractors

Entity extractors extract entities, such as person names or locations, from the user message.

note

If you use multiple entity extractors, we advise that each extractor targets an exclusive set of entity types. For example, use Duckling to extract dates and times, and DIETClassifier to extract person names. Otherwise, if multiple extractors target the same entity types, it is very likely that entities will be extracted multiple times.

For example, if you use two or more general purpose extractors like MitieEntityExtractor, DIETClassifier, or CRFEntityExtractor, the entity types in your training data will be found and extracted by all of them. If the slots you are filling with your entity types are of type text, then the last extractor in your pipeline will win. If the slot is of type list, then all results will be added to the list, including duplicates.

Another, less obvious case of duplicate/overlapping extraction can happen even if extractors focus on different entity types. Imagine a food delivery bot and a user message like I would like to order the Monday special. Hypothetically, if your time extractor's performance isn't very good, it might extract Monday here as a time for the order, and your other extractor might extract Monday special as the meal. If you struggle with overlapping entities of this sort, it might make sense to add additional training data to improve your extractor. If that does not suffice, you can add a custom component that resolves conflicts in entity extraction according to your own logic.

MitieEntityExtractor

  • Outputs

    entities

  • Output-Example

    {
    "entities": [{
    "value": "New York City",
    "start": 20,
    "end": 33,
    "confidence": null,
    "entity": "city",
    "extractor": "MitieEntityExtractor"
    }]
    }
  • Description

    MitieEntityExtractor uses the MITIE entity extraction to find entities in a message. The underlying classifier is using a multi class linear SVM with a sparse linear kernel and custom features. The MITIE component does not provide entity confidence values.

    note

    This entity extractor does not rely on any featurizer as it extracts features on its own.

  • Configuration

    pipeline:
    - name: "MitieEntityExtractor"

SpacyEntityExtractor

  • Short

    spaCy entity extraction

  • Outputs

    entities

  • Output-Example

    {
    "entities": [{
    "value": "New York City",
    "start": 20,
    "end": 33,
    "confidence": null,
    "entity": "city",
    "extractor": "SpacyEntityExtractor"
    }]
    }
  • Description

    Using spaCy this component predicts the entities of a message. spaCy uses a statistical BILOU transition model. As of now, this component can only use the spaCy builtin entity extraction models and can not be retrained. This extractor does not provide any confidence scores.

    You can test out spaCy's entity extraction models in this interactive demo. Note that some spaCy models are highly case-sensitive.

note

The SpacyEntityExtractor extractor does not provide a confidence level and will always return null.

  • Configuration

    Configure which dimensions, i.e. entity types, the spaCy component should extract. A full list of available dimensions can be found in the spaCy documentation. Leaving the dimensions option unspecified will extract all available dimensions.

    pipeline:
    - name: "SpacyEntityExtractor"
    # dimensions to extract
    dimensions: ["PERSON", "LOC", "ORG", "PRODUCT"]

CRFEntityExtractor

  • Short

    Conditional random field (CRF) entity extraction

  • Outputs

    entities

  • Requires

    tokens and dense_features (optional)

  • Output-Example

    {
    "entities": [{
    "value": "New York City",
    "start": 20,
    "end": 33,
    "entity": "city",
    "confidence": 0.874,
    "extractor": "CRFEntityExtractor"
    }]
    }
  • Description

    This component implements a conditional random fields (CRF) to do named entity recognition. CRFs can be thought of as an undirected Markov chain where the time steps are words and the states are entity classes. Features of the words (capitalization, POS tagging, etc.) give probabilities to certain entity classes, as are transitions between neighbouring entity tags: the most likely set of tags is then calculated and returned.

If you want to pass custom features, such as pre-trained word embeddings, to CRFEntityExtractor, you can add any dense featurizer to the pipeline before the CRFEntityExtractor and subsequently configure CRFEntityExtractor to make use of the dense features by adding "text_dense_feature" to its feature configuration. CRFEntityExtractor automatically finds the additional dense features and checks if the dense features are an iterable of len(tokens), where each entry is a vector. A warning will be shown in case the check fails. However, CRFEntityExtractor will continue to train just without the additional custom features. In case dense features are present, CRFEntityExtractor will pass the dense features to sklearn_crfsuite and use them for training.

  • Configuration

    CRFEntityExtractor has a list of default features to use. However, you can overwrite the default configuration. The following features are available:

    =================== ==========================================================================================
    Feature Name Description
    =================== ==========================================================================================
    low word identity - use the lower-cased token as a feature.
    upper Checks if the token is upper case.
    title Checks if the token starts with an uppercase character and all remaining characters are
    lowercased.
    digit Checks if the token contains just digits.
    prefix5 Take the first five characters of the token.
    prefix2 Take the first two characters of the token.
    suffix5 Take the last five characters of the token.
    suffix3 Take the last three characters of the token.
    suffix2 Take the last two characters of the token.
    suffix1 Take the last character of the token.
    pos Take the Part-of-Speech tag of the token (``SpacyTokenizer`` required).
    pos2 Take the first two characters of the Part-of-Speech tag of the token
    (``SpacyTokenizer`` required).
    pattern Take the patterns defined by ``RegexFeaturizer``.
    bias Add an additional "bias" feature to the list of features.
    text_dense_features Adds additional features from a dense featurizer.
    =================== ==========================================================================================

    As the featurizer is moving over the tokens in a user message with a sliding window, you can define features for previous tokens, the current token, and the next tokens in the sliding window. You define the features as [before, token, after] array.

    Additional you can set a flag to determine whether to use the BILOU tagging schema or not.

    • BILOU_flag determines whether to use BILOU tagging or not. Default True.
    pipeline:
    - name: "CRFEntityExtractor"
    # BILOU_flag determines whether to use BILOU tagging or not.
    "BILOU_flag": True
    # features to extract in the sliding window
    "features": [
    ["low", "title", "upper"],
    [
    "bias",
    "low",
    "prefix5",
    "prefix2",
    "suffix5",
    "suffix3",
    "suffix2",
    "upper",
    "title",
    "digit",
    "pattern",
    "text_dense_features"
    ],
    ["low", "title", "upper"],
    ]
    # The maximum number of iterations for optimization algorithms.
    "max_iterations": 50
    # weight of the L1 regularization
    "L1_c": 0.1
    # weight of the L2 regularization
    "L2_c": 0.1
    # Name of dense featurizers to use.
    # If list is empty all available dense features are used.
    "featurizers": []
    # Indicated whether a list of extracted entities should be split into individual entities for a given entity type
    "split_entities_by_comma":
    address: False
    email: True
    note

    If POS features are used (pos or pos2), you need to have SpacyTokenizer in your pipeline.

    note

    If pattern features are used, you need to have RegexFeaturizer in your pipeline.

    note

    If text_dense_features features are used, you need to have a dense featurizer (e.g. LanguageModelFeaturizer) in your pipeline.

DucklingEntityExtractor

  • Short

    Duckling lets you extract common entities like dates, amounts of money, distances, and others in a number of languages.

  • Outputs

    entities

  • Requires

    Nothing

  • Output-Example

    {
    "entities": [{
    "end": 53,
    "entity": "time",
    "start": 48,
    "value": "2017-04-10T00:00:00.000+02:00",
    "confidence": 1.0,
    "extractor": "DucklingEntityExtractor"
    }]
    }
  • Description

    To use this component you need to run a duckling server. The easiest option is to spin up a docker container using docker run -p 8000:8000 rasa/duckling.

    Alternatively, you can install duckling directly on your machine and start the server.

    Duckling allows to recognize dates, numbers, distances and other structured entities and normalizes them. Please be aware that duckling tries to extract as many entity types as possible without providing a ranking. For example, if you specify both number and time as dimensions for the duckling component, the component will extract two entities: 10 as a number and in 10 minutes as a time from the text I will be there in 10 minutes. In such a situation, your application would have to decide which entity type is be the correct one. The extractor will always return 1.0 as a confidence, as it is a rule based system.

    The list of supported languages can be found in the Duckling GitHub repository.

  • Configuration

    Configure which dimensions, i.e. entity types, the duckling component should extract. A full list of available dimensions can be found in the duckling project readme. Leaving the dimensions option unspecified will extract all available dimensions.

    pipeline:
    - name: "DucklingEntityExtractor"
    # url of the running duckling server
    url: "http://localhost:8000"
    # dimensions to extract
    dimensions: ["time", "number", "amount-of-money", "distance"]
    # allows you to configure the locale, by default the language is
    # used
    locale: "de_DE"
    # if not set the default timezone of Duckling is going to be used
    # needed to calculate dates from relative expressions like "tomorrow"
    timezone: "Europe/Berlin"
    # Timeout for receiving response from http url of the running duckling server
    # if not set the default timeout of duckling http url is set to 3 seconds.
    timeout : 3

DIETClassifier

  • Short

    Dual Intent Entity Transformer (DIET) used for intent classification and entity extraction

  • Description

    You can find the detailed description of the DIETClassifier under the section Intent Classifiers.

RegexEntityExtractor

  • Short

    Extracts entities using the lookup tables and/or regexes defined in the training data

  • Outputs

    entities

  • Requires

    Nothing

  • Description

    This component extract entities using the lookup tables and regexes defined in the training data. The component checks if the user message contains an entry of one of the lookup tables or matches one of the regexes. If a match is found, the value is extracted as entity.

    This component only uses those regex features that have a name equal to one of the entities defined in the training data. Make sure to annotate at least one example per entity.

    note

    When you use this extractor in combination with MitieEntityExtractor, CRFEntityExtractor, or DIETClassifier it can lead to multiple extraction of entities. Especially if many training sentences have entity annotations for the entity types for which you also have defined regexes. See the big info box at the start of the entity extractor section for more info on multiple extraction.

    In the case where you seem to need both this RegexEntityExtractor and another of the aforementioned statistical extractors, we advise you to consider one of the following two options.

    Option 1 is advisable when you have exclusive entity types for each type of extractor. To make the sure the extractors don't interfere with one another annotate only one example sentence for each regex/lookup entity type, but not more.

    Option 2 is useful when you want to use regexes matches as additional signal for your statistical extractor, but you don't have separate entity types. In this case you will want to 1) add the RegexFeaturizer before the extractors in your pipeline 2) annotate all your entity examples in the training data and 3) remove the RegexEntityExtractor from your pipeline. This way, your statistical extractors will receive additional signal about the presence of regex matches and will be able to statistically determine when to rely on these matches and when not to.

  • Configuration

    Make the entity extractor case sensitive by adding the case_sensitive: True option, the default being case_sensitive: False.

    To correctly process languages such as Chinese that don't use whitespace for word separation, the user needs to add the use_word_boundaries: False option, the default being use_word_boundaries: True.

    pipeline:
    - name: RegexEntityExtractor
    # text will be processed with case insensitive as default
    case_sensitive: False
    # use lookup tables to extract entities
    use_lookup_tables: True
    # use regexes to extract entities
    use_regexes: True
    # use match word boundaries for lookup table
    use_word_boundaries: True

EntitySynonymMapper

  • Short

    Maps synonymous entity values to the same value.

  • Outputs

    Modifies existing entities that previous entity extraction components found.

  • Description

    If the training data contains defined synonyms, this component will make sure that detected entity values will be mapped to the same value. For example, if your training data contains the following examples:

    [
    {
    "text": "I moved to New York City",
    "intent": "inform_relocation",
    "entities": [{
    "value": "nyc",
    "start": 11,
    "end": 24,
    "entity": "city",
    }]
    },
    {
    "text": "I got a new flat in NYC.",
    "intent": "inform_relocation",
    "entities": [{
    "value": "nyc",
    "start": 20,
    "end": 23,
    "entity": "city",
    }]
    }
    ]

    This component will allow you to map the entities New York City and NYC to nyc. The entity extraction will return nyc even though the message contains NYC. When this component changes an existing entity, it appends itself to the processor list of this entity.

  • Configuration

    pipeline:
    - name: "EntitySynonymMapper"
    note

    When using the EntitySynonymMapper as part of an NLU pipeline, it will need to be placed below any entity extractors in the configuration file.

Combined Intent Classifiers and Entity Extractors

DIETClassifier

  • Short

    Dual Intent Entity Transformer (DIET) used for intent classification and entity extraction

  • Outputs

    entities, intent and intent_ranking

  • Requires

    dense_features and/or sparse_features for user message and optionally the intent

  • Output-Example

    {
    "intent": {"name": "greet", "confidence": 0.7800},
    "intent_ranking": [
    {
    "confidence": 0.7800,
    "name": "greet"
    },
    {
    "confidence": 0.1400,
    "name": "goodbye"
    },
    {
    "confidence": 0.0800,
    "name": "restaurant_search"
    }
    ],
    "entities": [{
    "end": 53,
    "entity": "time",
    "start": 48,
    "value": "2017-04-10T00:00:00.000+02:00",
    "confidence": 1.0,
    "extractor": "DIETClassifier"
    }]
    }
  • Description

    DIET (Dual Intent and Entity Transformer) is a multi-task architecture for intent classification and entity recognition. The architecture is based on a transformer which is shared for both tasks. A sequence of entity labels is predicted through a Conditional Random Field (CRF) tagging layer on top of the transformer output sequence corresponding to the input sequence of tokens. For the intent labels the transformer output for the complete utterance and intent labels are embedded into a single semantic vector space. We use the dot-product loss to maximize the similarity with the target label and minimize similarities with negative samples.

    If you want to learn more about the model, check out the Algorithm Whiteboard series on YouTube, where we explain the model architecture in detail.

    note

    If during prediction time a message contains only words unseen during training and no Out-Of-Vocabulary preprocessor was used, an empty intent None is predicted with confidence 0.0. This might happen if you only use the CountVectorsFeaturizer with a word analyzer as featurizer. If you use the char_wb analyzer, you should always get an intent with a confidence value > 0.0.

  • Configuration

    If you want to use the DIETClassifier just for intent classification, set entity_recognition to False. If you want to do only entity recognition, set intent_classification to False. By default DIETClassifier does both, i.e. entity_recognition and intent_classification are set to True.

    You can define a number of hyperparameters to adapt the model. If you want to adapt your model, start by modifying the following parameters:

    • epochs: This parameter sets the number of times the algorithm will see the training data (default: 300). One epoch is equals to one forward pass and one backward pass of all the training examples. Sometimes the model needs more epochs to properly learn. Sometimes more epochs don't influence the performance. The lower the number of epochs the faster the model is trained.

    • hidden_layers_sizes: This parameter allows you to define the number of feed forward layers and their output dimensions for user messages and intents (default: text: [], label: []). Every entry in the list corresponds to a feed forward layer. For example, if you set text: [256, 128], we will add two feed forward layers in front of the transformer. The vectors of the input tokens (coming from the user message) will be passed on to those layers. The first layer will have an output dimension of 256 and the second layer will have an output dimension of 128. If an empty list is used (default behavior), no feed forward layer will be added. Make sure to use only positive integer values. Usually, numbers of power of two are used. Also, it is usual practice to have decreasing values in the list: next value is smaller or equal to the value before.

    • embedding_dimension: This parameter defines the output dimension of the embedding layers used inside the model (default: 20). We are using multiple embeddings layers inside the model architecture. For example, the vector of the complete utterance and the intent is passed on to an embedding layer before they are compared and the loss is calculated.

    • number_of_transformer_layers: This parameter sets the number of transformer layers to use (default: 2). The number of transformer layers corresponds to the transformer blocks to use for the model.

    • transformer_size: This parameter sets the number of units in the transformer (default: 256). The vectors coming out of the transformers will have the given transformer_size. The transformer_size should be a multiple of the number_of_attention_heads parameter, the training exits with an error otherwise.

    • connection_density: This parameter defines the fraction of kernel weights that are set to non zero values for all feed forward layers in the model (default: 0.2). The value should be between 0 and 1. If you set connection_density to 1, no kernel weights will be set to 0, the layer acts as a standard feed forward layer. You should not set connection_density to 0 as this would result in all kernel weights being 0, i.e. the model is not able to learn.

    • BILOU_flag: This parameter determines whether to use BILOU tagging or not. Default True.

    The above configuration parameters are the ones you should configure to fit your model to your data. However, additional parameters exist that can be adapted.

    More configurable parameters
    +---------------------------------+------------------+--------------------------------------------------------------+
    | Parameter | Default Value | Description |
    +=================================+==================+==============================================================+
    | hidden_layers_sizes | text: [] | Hidden layer sizes for layers before the embedding layers |
    | | label: [] | for user messages and labels. The number of hidden layers is |
    | | | equal to the length of the corresponding. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | share_hidden_layers | False | Whether to share the hidden layer weights between user |
    | | | messages and labels. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | transformer_size | 256 | Number of units in transformer. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | number_of_transformer_layers | 2 | Number of transformer layers. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | number_of_attention_heads | 4 | Number of attention heads in transformer. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | use_key_relative_attention | False | If 'True' use key relative embeddings in attention. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | use_value_relative_attention | False | If 'True' use value relative embeddings in attention. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | max_relative_position | None | Maximum position for relative embeddings. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | unidirectional_encoder | False | Use a unidirectional or bidirectional encoder. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | batch_size | [64, 256] | Initial and final value for batch sizes. |
    | | | Batch size will be linearly increased for each epoch. |
    | | | If constant `batch_size` is required, pass an int, e.g. `8`. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | batch_strategy | "balanced" | Strategy used when creating batches. |
    | | | Can be either 'sequence' or 'balanced'. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | epochs | 300 | Number of epochs to train. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | random_seed | None | Set random seed to any 'int' to get reproducible results. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | learning_rate | 0.001 | Initial learning rate for the optimizer. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | embedding_dimension | 20 | Dimension size of embedding vectors. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | dense_dimension | text: 128 | Dense dimension for sparse features to use if no dense |
    | | label: 20 | features are present. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | concat_dimension | text: 128 | Concat dimension for sequence and sentence features. |
    | | label: 20 | |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | number_of_negative_examples | 20 | The number of incorrect labels. The algorithm will minimize |
    | | | their similarity to the user input during training. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | similarity_type | "auto" | Type of similarity measure to use, either 'auto' or 'cosine' |
    | | | or 'inner'. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | loss_type | "cross_entropy" | The type of the loss function, either 'cross_entropy' |
    | | | or 'margin'. Type 'margin' is only compatible with |
    | | | "model_confidence=cosine", |
    | | | which is deprecated (see changelog for 2.3.4). |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | ranking_length | 10 | Number of top intents to report. Set to 0 to report all |
    | | | intents. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | renormalize_confidences | False | Normalize the reported top intents. Applicable only with loss|
    | | | type 'cross_entropy' and 'softmax' confidences. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | maximum_positive_similarity | 0.8 | Indicates how similar the algorithm should try to make |
    | | | embedding vectors for correct labels. |
    | | | Should be 0.0 < ... < 1.0 for 'cosine' similarity type. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | maximum_negative_similarity | -0.4 | Maximum negative similarity for incorrect labels. |
    | | | Should be -1.0 < ... < 1.0 for 'cosine' similarity type. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | use_maximum_negative_similarity | True | If 'True' the algorithm only minimizes maximum similarity |
    | | | over incorrect intent labels, used only if 'loss_type' is |
    | | | set to 'margin'. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | scale_loss | False | Scale loss inverse proportionally to confidence of correct |
    | | | prediction. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | regularization_constant | 0.002 | The scale of regularization. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | negative_margin_scale | 0.8 | The scale of how important it is to minimize the maximum |
    | | | similarity between embeddings of different labels. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | connection_density | 0.2 | Connection density of the weights in dense layers. |
    | | | Value should be between 0 and 1. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | drop_rate | 0.2 | Dropout rate for encoder. Value should be between 0 and 1. |
    | | | The higher the value the higher the regularization effect. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | drop_rate_attention | 0.0 | Dropout rate for attention. Value should be between 0 and 1. |
    | | | The higher the value the higher the regularization effect. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | use_sparse_input_dropout | True | If 'True' apply dropout to sparse input tensors. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | use_dense_input_dropout | True | If 'True' apply dropout to dense input tensors. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | evaluate_every_number_of_epochs | 20 | How often to calculate validation accuracy. |
    | | | Set to '-1' to evaluate just once at the end of training. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | evaluate_on_number_of_examples | 0 | How many examples to use for hold out validation set. |
    | | | Large values may hurt performance, e.g. model accuracy. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | intent_classification | True | If 'True' intent classification is trained and intents are |
    | | | predicted. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | entity_recognition | True | If 'True' entity recognition is trained and entities are |
    | | | extracted. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | use_masked_language_model | False | If 'True' random tokens of the input message will be masked |
    | | | and the model has to predict those tokens. It acts like a |
    | | | regularizer and should help to learn a better contextual |
    | | | representation of the input. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | BILOU_flag | True | If 'True', additional BILOU tags are added to entity labels. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | tensorboard_log_directory | None | If you want to use tensorboard to visualize training |
    | | | metrics, set this option to a valid output directory. You |
    | | | can view the training metrics after training in tensorboard |
    | | | via 'tensorboard --logdir <path-to-given-directory>'. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | tensorboard_log_level | "epoch" | Define when training metrics for tensorboard should be |
    | | | logged. Either after every epoch ('epoch') or for every |
    | | | training step ('batch'). |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | featurizers | [] | List of featurizer names (alias names). Only features |
    | | | coming from the listed names are used. If list is empty |
    | | | all available features are used. |
    +---------------------------------+------------------+--------------------------------------------------------------+
    | checkpoint_model | False | Save the best performing model during training. Models are |
    | | | stored to the location specified by `--out`. Only the one |
    | | | best model will be saved. |
    | | | Requires `evaluate_on_number_of_examples > 0` and |
    | | | `evaluate_every_number_of_epochs > 0` |
    +---------------------------------+------------------+--------------------------------------------------------------+
    note

    Parameter maximum_negative_similarity is set to a negative value to mimic the original starspace algorithm in the case maximum_negative_similarity = maximum_positive_similarity and use_maximum_negative_similarity = False. See starspace paper for details.

Selectors

Selectors predict a bot response from a set of candidate responses.

ResponseSelector

  • Short

    Response Selector

  • Outputs

    A dictionary with the key as the retrieval intent of the response selector and value containing predicted responses, confidence and the response key under the retrieval intent

  • Requires

    dense_features and/or sparse_features for user messages and response

  • Output-Example

    The parsed output from NLU will have a property named response_selector containing the output for each response selector component. Each response selector is identified by retrieval_intent parameter of that response selector and stores two properties:

    • response: The predicted response key under the corresponding retrieval intent, prediction's confidence and the associated responses.

    • ranking: Ranking with confidences of top 10 candidate response keys.

    Example result:

    {
    "response_selector": {
    "faq": {
    "response": {
    "id": 1388783286124361986,
    "confidence": 0.7,
    "intent_response_key": "chitchat/ask_weather",
    "responses": [
    {
    "text": "It's sunny in Berlin today",
    "image": "https://i.imgur.com/nGF1K8f.jpg"
    },
    {
    "text": "I think it's about to rain."
    }
    ],
    "utter_action": "utter_chitchat/ask_weather"
    },
    "ranking": [
    {
    "id": 1388783286124361986,
    "confidence": 0.7,
    "intent_response_key": "chitchat/ask_weather"
    },
    {
    "id": 1388783286124361986,
    "confidence": 0.3,
    "intent_response_key": "chitchat/ask_name"
    }
    ]
    }
    }
    }

    If the retrieval_intent parameter of a particular response selector was left to its default value, the corresponding response selector will be identified as default in the returned output.

    {
    "response_selector": {
    "default": {
    "response": {
    "id": 1388783286124361986,
    "confidence": 0.7,
    "intent_response_key": "chitchat/ask_weather",
    "responses": [
    {
    "text": "It's sunny in Berlin today",
    "image": "https://i.imgur.com/nGF1K8f.jpg"
    },
    {
    "text": "I think it's about to rain."
    }
    ],
    "utter_action": "utter_chitchat/ask_weather"
    },
    "ranking": [
    {
    "id": 1388783286124361986,
    "confidence": 0.7,
    "intent_response_key": "chitchat/ask_weather"
    },
    {
    "id": 1388783286124361986,
    "confidence": 0.3,
    "intent_response_key": "chitchat/ask_name"
    }
    ]
    }
    }
    }
  • Description

    Response Selector component can be used to build a response retrieval model to directly predict a bot response from a set of candidate responses. The prediction of this model is used by the dialogue manager to utter the predicted responses. It embeds user inputs and response labels into the same space and follows the exact same neural network architecture and optimization as the DIETClassifier.

    To use this component, your training data should contain retrieval intents. To define these, checkout documentation on NLU training examples and documentation on defining response utterances for retrieval intents.

    note

    If during prediction time a message contains only words unseen during training and no Out-Of-Vocabulary preprocessor was used, an empty response None is predicted with confidence 0.0. This might happen if you only use the CountVectorsFeaturizer with a word analyzer as featurizer. If you use the char_wb analyzer, you should always get a response with a confidence value > 0.0.

  • Configuration

    The algorithm includes almost all the hyperparameters that DIETClassifier uses. If you want to adapt your model, start by modifying the following parameters:

    • epochs: This parameter sets the number of times the algorithm will see the training data (default: 300). One epoch is equals to one forward pass and one backward pass of all the training examples. Sometimes the model needs more epochs to properly learn. Sometimes more epochs don't influence the performance. The lower the number of epochs the faster the model is trained.

    • hidden_layers_sizes: This parameter allows you to define the number of feed forward layers and their output dimensions for user messages and intents (default: text: [256, 128], label: [256, 128]). Every entry in the list corresponds to a feed forward layer. For example, if you set text: [256, 128], we will add two feed forward layers in front of the transformer. The vectors of the input tokens (coming from the user message) will be passed on to those layers. The first layer will have an output dimension of 256 and the second layer will have an output dimension of 128. If an empty list is used (default behavior), no feed forward layer will be added. Make sure to use only positive integer values. Usually, numbers of power of two are used. Also, it is usual practice to have decreasing values in the list: next value is smaller or equal to the value before.

    • embedding_dimension: This parameter defines the output dimension of the embedding layers used inside the model (default: 20). We are using multiple embeddings layers inside the model architecture. For example, the vector of the complete utterance and the intent is passed on to an embedding layer before they are compared and the loss is calculated.

    • number_of_transformer_layers: This parameter sets the number of transformer layers to use (default: 0). The number of transformer layers corresponds to the transformer blocks to use for the model.

    • transformer_size: This parameter sets the number of units in the transformer (default: None). The vectors coming out of the transformers will have the given transformer_size. The transformer_size should be a multiple of the number_of_attention_heads parameter, the training exits with an error otherwise.

    • connection_density: This parameter defines the fraction of kernel weights that are set to non zero values for all feed forward layers in the model (default: 0.2). The value should be between 0 and 1. If you set connection_density to 1, no kernel weights will be set to 0, the layer acts as a standard feed forward layer. You should not set connection_density to 0 as this would result in all kernel weights being 0, i.e. the model is not able to learn.

    • constrain_similarities: This parameter when set to True applies a sigmoid cross entropy loss over all similarity terms. This helps in keeping similarities between input and negative labels to smaller values. This should help in better generalization of the model to real world test sets.

    • model_confidence: This parameter allows the user to configure how confidences are computed during inference. It can take only one value as input which is softmax. In softmax, confidences are in the range [0, 1]. The computed similarities are normalized with the softmax activation function.

    The component can also be configured to train a response selector for a particular retrieval intent. The parameter retrieval_intent sets the name of the retrieval intent for which this response selector model is trained. Default is None, i.e. the model is trained for all retrieval intents.

    In its default configuration, the component uses the retrieval intent with the response key(e.g. faq/ask_name) as the label for training. Alternatively, it can also be configured to use the text of the responses as the training label by switching use_text_as_label to True. In this mode, the component will use the first available response which has a text attribute for training. If none are found, it falls back to using the retrieval intent combined with the response key as the label.

    examples and tutorials

    Check out the responseselectorbot for an example of how you can use the ResponseSelector component in your assistant. Additionally, you will find this tutorial on handling FAQs using a ResponseSelector useful as well.

The above configuration parameters are the ones you should configure to fit your model to your data. However, additional parameters exist that can be adapted.

More configurable parameters
+---------------------------------+-------------------+--------------------------------------------------------------+
| Parameter | Default Value | Description |
+=================================+===================+==============================================================+
| hidden_layers_sizes | text: [256, 128] | Hidden layer sizes for layers before the embedding layers |
| | label: [256, 128] | for user messages and labels. The number of hidden layers is |
| | | equal to the length of the corresponding list. We recommend |
| | | disabling the hidden layers (by providing empty lists) when |
| | | the transformer is enabled. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| share_hidden_layers | False | Whether to share the hidden layer weights between user |
| | | messages and labels. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| transformer_size | None | Number of units in the transformer. When a positive value is |
| | | provided for `number_of_transformer_layers`, the default size|
| | | becomes `256`. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| number_of_transformer_layers | 0 | Number of transformer layers; positive values enable the |
| | | transformer. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| number_of_attention_heads | 4 | Number of attention heads in transformer. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| use_key_relative_attention | False | If 'True' use key relative embeddings in attention. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| use_value_relative_attention | False | If 'True' use value relative embeddings in attention. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| max_relative_position | None | Maximum position for relative embeddings. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| unidirectional_encoder | False | Use a unidirectional or bidirectional encoder. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| batch_size | [64, 256] | Initial and final value for batch sizes. |
| | | Batch size will be linearly increased for each epoch. |
| | | If constant `batch_size` is required, pass an int, e.g. `8`. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| batch_strategy | "balanced" | Strategy used when creating batches. |
| | | Can be either 'sequence' or 'balanced'. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| epochs | 300 | Number of epochs to train. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| random_seed | None | Set random seed to any 'int' to get reproducible results. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| learning_rate | 0.001 | Initial learning rate for the optimizer. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| embedding_dimension | 20 | Dimension size of embedding vectors. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| dense_dimension | text: 512 | Dense dimension for sparse features to use if no dense |
| | label: 512 | features are present. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| concat_dimension | text: 512 | Concat dimension for sequence and sentence features. |
| | label: 512 | |
+---------------------------------+-------------------+--------------------------------------------------------------+
| number_of_negative_examples | 20 | The number of incorrect labels. The algorithm will minimize |
| | | their similarity to the user input during training. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| similarity_type | "auto" | Type of similarity measure to use, either 'auto' or 'cosine' |
| | | or 'inner'. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| loss_type | "cross_entropy" | The type of the loss function, either 'cross_entropy' |
| | | or 'margin'. Type 'margin' is only compatible with |
| | | "model_confidence=cosine", |
| | | which is deprecated (see changelog for 2.3.4). |
+---------------------------------+-------------------+--------------------------------------------------------------+
| ranking_length | 10 | Number of top responses to report. Set to 0 to report all |
| | | responses. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| renormalize_confidences | False | Normalize the top responses. Applicable only with loss type |
| | | 'cross_entropy' and 'softmax' confidences. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| maximum_positive_similarity | 0.8 | Indicates how similar the algorithm should try to make |
| | | embedding vectors for correct labels. |
| | | Should be 0.0 < ... < 1.0 for 'cosine' similarity type. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| maximum_negative_similarity | -0.4 | Maximum negative similarity for incorrect labels. |
| | | Should be -1.0 < ... < 1.0 for 'cosine' similarity type. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| use_maximum_negative_similarity | True | If 'True' the algorithm only minimizes maximum similarity |
| | | over incorrect intent labels, used only if 'loss_type' is |
| | | set to 'margin'. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| scale_loss | True | Scale loss inverse proportionally to confidence of correct |
| | | prediction. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| regularization_constant | 0.002 | The scale of regularization. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| negative_margin_scale | 0.8 | The scale of how important is to minimize the maximum |
| | | similarity between embeddings of different labels. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| connection_density | 0.2 | Connection density of the weights in dense layers. |
| | | Value should be between 0 and 1. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| drop_rate | 0.2 | Dropout rate for encoder. Value should be between 0 and 1. |
| | | The higher the value the higher the regularization effect. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| drop_rate_attention | 0.0 | Dropout rate for attention. Value should be between 0 and 1. |
| | | The higher the value the higher the regularization effect. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| use_sparse_input_dropout | False | If 'True' apply dropout to sparse input tensors. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| use_dense_input_dropout | False | If 'True' apply dropout to dense input tensors. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| evaluate_every_number_of_epochs | 20 | How often to calculate validation accuracy. |
| | | Set to '-1' to evaluate just once at the end of training. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| evaluate_on_number_of_examples | 0 | How many examples to use for hold out validation set. |
| | | Large values may hurt performance, e.g. model accuracy. |
| | | Set to 0 for no validation. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| use_masked_language_model | False | If 'True' random tokens of the input message will be masked |
| | | and the model should predict those tokens. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| retrieval_intent | None | Name of the intent for which this response selector model is |
| | | trained. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| use_text_as_label | False | Whether to use the actual text of the response as the label |
| | | for training the response selector. Otherwise, it uses the |
| | | response key as the label. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| tensorboard_log_directory | None | If you want to use tensorboard to visualize training |
| | | metrics, set this option to a valid output directory. You |
| | | can view the training metrics after training in tensorboard |
| | | via 'tensorboard --logdir <path-to-given-directory>'. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| tensorboard_log_level | "epoch" | Define when training metrics for tensorboard should be |
| | | logged. Either after every epoch ("epoch") or for every |
| | | training step ("batch"). |
+---------------------------------+-------------------+--------------------------------------------------------------+
| featurizers | [] | List of featurizer names (alias names). Only features |
| | | coming from the listed names are used. If list is empty |
| | | all available features are used. |
+---------------------------------+-------------------+--------------------------------------------------------------+
| checkpoint_model | False | Save the best performing model during training. Models are |
| | | stored to the location specified by `--out`. Only the one |
| | | best model will be saved. |
| | | Requires `evaluate_on_number_of_examples > 0` and |
| | | `evaluate_every_number_of_epochs > 0` |
+---------------------------------+-------------------+--------------------------------------------------------------+
| constrain_similarities | False | If `True`, applies sigmoid on all similarity terms and adds |
| | | it to the loss function to ensure that similarity values are |
| | | approximately bounded. Used only if `loss_type=cross_entropy`|
+---------------------------------+-------------------+--------------------------------------------------------------+
| model_confidence | "softmax" | Affects how model's confidence for each response label |
| | | is computed. Currently, only one value is supported: |
| | | 1. `softmax` - Similarities between input and response label |
| | | embeddings are post-processed with a softmax function, |
| | | as a result of which confidence for all labels sum up to 1. |
+---------------------------------+-------------------+--------------------------------------------------------------+
note

Parameter maximum_negative_similarity is set to a negative value to mimic the original starspace algorithm in the case maximum_negative_similarity = maximum_positive_similarity and use_maximum_negative_similarity = False. See starspace paper for details.

Custom Components

New in 3.0

Rasa 3.0 unified the implementation of NLU components and policies. This requires changes to custom components written for earlier versions of Rasa Open Source. Please see the migration guide for a step-by-step guide for the migration.

You can create a custom component to perform a specific task which NLU doesn't currently offer (for example, sentiment analysis).

You can add a custom component to your pipeline by adding the module path. So if you have a module called sentiment containing a SentimentAnalyzer class:

pipeline:
- name: "sentiment.SentimentAnalyzer"

See the guide on custom graph components for a complete guide on custom components. Also be sure to read the section on the Component Lifecycle.


  1. Note that model_confidence: cosineis deprecated in 2.3.4 (see changelog) and cannot be specified in the config, however model_confidence: cosine will be used regardless of the config if loss_type: margin is specified.