mirror of
https://github.com/explosion/spaCy.git
synced 2024-09-21 19:39:13 +03:00
51 lines
2.8 KiB
Plaintext
51 lines
2.8 KiB
Plaintext
//- 💫 DOCS > USAGE > SPACY 101 > TRAINING
|
||
|
||
p
|
||
| spaCy's models are #[strong statistical] and every "decision" they make –
|
||
| for example, which part-of-speech tag to assign, or whether a word is a
|
||
| named entity – is a #[strong prediction]. This prediction is based
|
||
| on the examples the model has seen during #[strong training]. To train
|
||
| a model, you first need training data – examples of text, and the
|
||
| labels you want the model to predict. This could be a part-of-speech tag,
|
||
| a named entity or any other information.
|
||
|
||
p
|
||
| The model is then shown the unlabelled text and will make a prediction.
|
||
| Because we know the correct answer, we can give the model feedback on its
|
||
| prediction in the form of an #[strong error gradient] of the
|
||
| #[strong loss function] that calculates the difference between the training
|
||
| example and the expected output. The greater the difference, the more
|
||
| significant the gradient and the updates to our model.
|
||
|
||
+aside
|
||
| #[strong Training data:] Examples and their annotations.#[br]
|
||
| #[strong Text:] The input text the model should predict a label for.#[br]
|
||
| #[strong Label:] The label the model should predict.#[br]
|
||
| #[strong Gradient:] Gradient of the loss function calculating the
|
||
| difference between input and expected output.
|
||
|
||
+graphic("/assets/img/training.svg")
|
||
include ../../assets/img/training.svg
|
||
|
||
p
|
||
| When training a model, we don't just want it to memorise our examples –
|
||
| we want it to come up with theory that can be
|
||
| #[strong generalised across other examples]. After all, we don't just want
|
||
| the model to learn that this one instance of "Amazon" right here is a
|
||
| company – we want it to learn that "Amazon", in contexts #[em like this],
|
||
| is most likely a company. That's why the training data should always be
|
||
| representative of the data we want to process. A model trained on
|
||
| Wikipedia, where sentences in the first person are extremely rare, will
|
||
| likely perform badly on Twitter. Similarly, a model trained on romantic
|
||
| novels will likely perform badly on legal text.
|
||
|
||
p
|
||
| This also means that in order to know how the model is performing,
|
||
| and whether it's learning the right things, you don't only need
|
||
| #[strong training data] – you'll also need #[strong evaluation data]. If
|
||
| you only test the model with the data it was trained on, you'll have no
|
||
| idea how well it's generalising. If you want to train a model from scratch,
|
||
| you usually need at least a few hundred examples for both training and
|
||
| evaluation. To update an existing model, you can already achieve decent
|
||
| results with very few examples – as long as they're representative.
|