Custom NER training basics in NLP - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the library needed for Named Entity Recognition (NER) training.

NLP

import [1]

Drag options to blanks, or click blank then click option'

Aspacy

Bnumpy

Cmatplotlib

Dpandas

Attempts:

3 left

💡 Hint

Common Mistakes

Importing unrelated libraries like numpy or pandas.

Forgetting to import the NLP library before training.

We use spacy for training custom Named Entity Recognition models.

2fill in blank

medium

Complete the code to load a blank English model for NER training.

NLP

nlp = spacy.blank('[1]')

Drag options to blanks, or click blank then click option'

Afr

Ben

Cde

Des

Attempts:

3 left

💡 Hint

Common Mistakes

Using other language codes like 'fr' or 'de' when training English NER.

Passing full language names instead of codes.

We use 'en' to load a blank English model for training.

3fill in blank

hard

Fix the error in adding the NER pipeline component to the model.

NLP

ner = nlp.add_pipe('[1]')

Drag options to blanks, or click blank then click option'

Atextcat

Btagger

Cparser

Dner

Attempts:

3 left

💡 Hint

Common Mistakes

Using unrelated pipeline components like 'textcat' or 'parser'.

Misspelling the component name.

The pipeline component for Named Entity Recognition is called 'ner'.

4fill in blank

hard

Fill both blanks to add a new entity label and prepare the optimizer.

NLP

ner.add_label('[1]')
optimizer = nlp.[2]()

Drag options to blanks, or click blank then click option'

APERSON

Btrain

Cbegin_training

DORG

Attempts:

3 left

💡 Hint

Common Mistakes

Using lowercase or unrelated strings as labels.

Calling a non-existent method like 'train()' instead of 'begin_training()'.

We add the entity label 'PERSON' and start training with begin_training().

5fill in blank

hard

Fill all three blanks to update the model with training data and print the loss.

NLP

for text, annotations in TRAIN_DATA:
    doc = nlp.make_doc(text)
    example = spacy.training.Example.from_dict(doc, [1])
    nlp.update([example], sgd=[2], losses=[3])

Drag options to blanks, or click blank then click option'

Aannotations

Boptimizer

Closses

Dtext

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the wrong variable for annotations or optimizer.

Not tracking losses correctly.

We use annotations for labels, optimizer for training, and losses to track errors.