Solution: Text Classification with spaCy

import random
import spacy
from spacy.training import Example
from spacy.pipeline.textcat import DEFAULT_SINGLE_TEXTCAT_MODEL
nlp = spacy.load("en_core_web_md")
# Define the training data
train_data = [
    ("Biden administration announces new COVID-19 vaccine mandate for federal workers", {"cats": {"REAL": 1, "FAKE": 0}}),
    ("Scientists discover new species of dinosaur in Argentina", {"cats": {"REAL": 1, "FAKE": 0}}),
    ("Stock market reaches record high after positive economic data", {"cats": {"REAL": 1, "FAKE": 0}}),
    ("COVID-19 vaccine causes infertility, clickbait article shows", {"cats": {"REAL": 0, "FAKE": 1}}),
    ("Donald Trump declares martial law and orders military takeover", {"cats": {"REAL": 0, "FAKE": 1}}),
    ("Hillary Clinton linked to human trafficking ring", {"cats": {"REAL": 0, "FAKE": 1}})
]
config = {
    "threshold": 0.5,
    "model": DEFAULT_SINGLE_TEXTCAT_MODEL
}
textcat = nlp.add_pipe("textcat", config=config)
textcat.add_label("REAL")
textcat.add_label("FAKE")
train_examples = [Example.from_dict(nlp.make_doc(text), label) for text,label in train_data]
textcat.initialize(lambda: train_examples, nlp=nlp)
epochs=20
with nlp.select_pipes(enable="textcat"):
    optimizer = nlp.resume_training()
    for i in range(epochs):
        random.shuffle(train_data)
        for text, label in train_data:
            doc = nlp.make_doc(text)
            example = Example.from_dict(doc, label)
            nlp.update([example], sgd=optimizer)
# Test the model
texts = [
    "According to a clickbait article COVID-19 vaccine causes infertility",
    "Covid-19 vaccine mandatory at work places"
]
for text in texts:
    doc = nlp(text)
    print(text, doc.cats)