Sentiment Analysis with Hugging Face

The pipeline shortcut

Hugging Face pipeline hides all the complexity of tokenization and model loading.

from transformers import pipeline

classifier = pipeline("sentiment-analysis")
print(classifier("I absolutely loved this tutorial!"))
# [{'label': 'POSITIVE', 'score': 0.9998}]

Beyond the default

Swap in any model from the Hub for your task or language:

pipeline("sentiment-analysis", model="distilbert-base-uncased-finetuned-sst-2-english")

The same pipeline interface works for translation, summarization, NER, and more — a great way to start before you fine-tune your own.