Warning: This document is for an old version of rasa NLU. The latest version is 0.11.3.


The evaluation script evaluate.py allows you to test your models performance for intent classification and entity recognition. You invoke this script supplying test data, model, and config file arguments:

python -m rasa_nlu.evaluate -d data/my_test.json -m models/my_model -c my_nlu_config.json

If you would like to evaluate your pipeline using crossvalidation, you can run the evaluation script with the mode crossvalidation flag. This gives you an estimate of how accurately a predictive model will perform in practice. Note that you cannot specify a model in this mode, as a new model will be trained on part of the data for every crossvalidation loop. An example invocation of your script would be:

python -m rasa_nlu.evaluate -d data/examples/rasa/demo-rasa.json -c sample_configs/config_spacy.json --mode crossvalidation

Intent Classification

The evaluation script will log precision, recall, and f1 measure for each intent and once summarized for all. Furthermore, it creates a confusion matrix for you to see which intents are mistaken for which others.

Entity Extraction

For each entity extractor, the evaluation script logs its performance per entity type in your training data. So if you use ner_crf and ner_duckling in your pipeline, it will log two evaluation tables containing recall, precision, and f1 measure for each entity type.

In the case ner_duckling we actually run the evaluation for each defined duckling dimension. If you use the time and ordinal dimensions, you would get two evaluation tables: one for ner_duckling (Time) and one for ner_duckling (Ordinal).

ner_synonyms does not create an evaluation table, because it only changes the value of the found entities and does not find entity boundaries itself.

Finally, keep in mind that entity types in your testing data have to match the output of the extraction components. This is particularly important for ner_duckling, because it is not fitted to your training data.

Entity Scoring

To evaluate entity extraction we apply a simple tag-based approach. We don’t consider BILOU tags, but only the entity type tags on a per token basis. For location entity like “near Alexanderplatz” we expect the labels “LOC” “LOC” instead of the BILOU-based “B-LOC” “L-LOC”. Our approach is more lenient when it comes to evaluation, as it rewards partial extraction and does not punish the splitting of entities. For example, the given the aforementioned entity “near Alexanderplatz” and a system that extracts “Alexanderplatz”, this reward the extraction of “Alexanderplatz” and punish the missed out word “near”. The BILOU-based approach, however, would label this as a complete failure since it expects Alexanderplatz to be labeled as a last token in an entity (L-LOC) instead of a single token entity (U-LOC). Also note, a splitted extraction of “near” and “Alexanderplatz” would get full scores on our approach and zero on the BILOU-based one.

Here’s a comparison between both different scoring mechanisms for the phrase “near Alexanderplatz tonight”:

extracted Simple tags (score) BILOU tags (score)
[near Alexanderplatz](loc) [tonight](time) loc loc time (3) B-loc L-loc U-time (3)
[near](loc) [Alexanderplatz](loc) [tonight](time) loc loc time (3) U-loc U-loc U-time (1)
near [Alexanderplatz](loc) [tonight](time) O loc time (2) O U-loc U-time (1)
[near](loc) Alexanderplatz [tonight](time) loc O time (2) U-loc O U-time (1)
[near Alexanderplatz tonight](loc) loc loc loc (2) B-loc I-loc L-loc (1)