Working with Hugging Face

Working with Hugging Face

What is Hugging Face?

Hugging Face is an open-source platform and community that provides state-of-the-art tools and models for Natural Language Processing (NLP) and other machine learning domains like computer vision, audio, and multimodal AI.

It’s best known for:

The Transformers library
Pretrained Large Language Models (LLMs) (like BERT, GPT-2, T5, etc.)
Hosting and sharing models via the Hugging Face Hub
No-code tools and APIs for AI applications

Why Use Hugging Face?

Quick start with pre-trained models
Efficient fine-tuning on your own data
Community-powered with thousands of open models
Compatible with TensorFlow, PyTorch, JAX

Key Libraries in Hugging Face

Library	Purpose
transformers	Pretrained models for NLP, vision, audio
datasets	Access and prepare thousands of ML datasets
huggingface_hub	Upload/download models, datasets, and spaces
gradio	Build and share interactive demos (Spaces)
evaluate	Standardized ML evaluation metrics

Getting Started with Hugging Face Transformers

Installation

pip install transformers

pip install datasets

Load a Pretrained Model

from transformers import pipeline

# Create a text classification pipeline

classifier = pipeline(“sentiment-analysis”)

# Use the model

result = classifier(“I love learning with Hugging Face!”)

print(result)

Output

[{‘label’: ‘POSITIVE’, ‘score’: 0.9998}]

Load a Model & Tokenizer Manually

from transformers import AutoTokenizer, AutoModelForSequenceClassification

model_name = “distilbert-base-uncased-finetuned-sst-2-english”

tokenizer = AutoTokenizer.from_pretrained(model_name)

model = AutoModelForSequenceClassification.from_pretrained(model_name)

Explore Datasets

from datasets import load_dataset

dataset = load_dataset(“imdb”)

print(dataset[“train”][0])

Using Hugging Face Hub

Search for Models & Datasets: Visit https://huggingface.co/models

You can search by:

Task (e.g., summarization, translation)
Framework (PyTorch, TensorFlow)
License, language, size, and more

Upload Your Own Model

huggingface-cli login

Then:

from huggingface_hub import HfApi

api = HfApi()

api.create_repo(name=”my-model”)

Common Use Cases

Use Case	Model/Pipeline
Sentiment Analysis	pipeline(“sentiment-analysis”)
Text Summarization	pipeline(“summarization”)
Question Answering	pipeline(“question-answering”)
Translation	pipeline(“translation”)
Text Generation	pipeline(“text-generation”)

Bonus: Hugging Face Spaces

Spaces lets you deploy and share ML-powered web apps using Gradio or Streamlit.

Try it: https://huggingface.co/spaces

Learning Outcomes

By the end of this module, you will:

Understand the Hugging Face ecosystem
Know how to use pre-trained models in your projects
Load and fine-tune datasets
Deploy ML demos with minimal code

Generative AI Mastery Quiz: From Basics to Real-World LLM Applications

Working with Hugging Face

Search

Login with your site account

Register a new account

Modal title