ner_training.py

import random
import spacy
from spacy.training.example import Example
from spacy.util import minibatch

# Initialize or load an NLP object and get the NER pipeline
nlp = spacy.blank("en")
nlp.add_pipe("ner")

# Initialize the optimizer
optimizer = nlp.begin_training()

# Train the model
for i in range(100):  # Number of epochs
    random.shuffle(TRAIN_DATA)
    losses = {}
    batches = minibatch(TRAIN_DATA, size=8)
    for batch in batches:
        texts, annotations = zip(*batch)
        example = []
        for i in range(len(texts)):
            doc = nlp.make_doc(texts[i])
            example.append(Example.from_dict(doc, annotations[i]))
        nlp.update(
            example,
            drop=0.5,  # Dropout rate
            losses=losses,
        )
    print(losses)