GoldParse

Motivating Model Training

• Spacy's models are statistical and every decision they make is a prediction
• This prediction is based on the examples the model has seen during training
• To train a model, we first need training data
• This could be a part-of-speech tag, a named entity, or any other information

Training with Annotations

• The GoldParse object is a collection of specified annotations for some given training examples
• These annotated training examples are called the gold standard
• A GoldParse object is initialized using the Doc object
• Specifically, the GoldParse object refers to the labels for some unlabeled training data
• Here's an example of a simple GoldParse for part-of-speech tags:

>>> vocab = Vocab(tag_map={"N": {"pos": "NOUN"},
...                        "V": {"pos": "VERB"}})
>>> doc = Doc(vocab, words=["I", "like", "stuff"])
>>> gold = GoldParse(doc, tags=["N", "V", "N"])

References

• Basics of Training Spacy Models
• GoldParse Class