Skip to content
  • Sign In
  • Try Now
View all events
Natural Language Processing

Hands-on Hugging Face for Natural Language Processing

Published by O'Reilly Media, Inc.

Intermediate content levelIntermediate

Leverage pretrained transformer models, data, and more from the Hugging Face platform

Course outcomes

  • Use transformer models with Hugging Face pipelines for your own applications
  • Configure Hugging Face tokenizers to preprocess text data
  • Fine-tune Hugging Face models for natural language processing tasks

Course description

The demand is growing for skilled NLP professionals who can navigate the complexities of language data and transform it into actionable insights or innovative applications. Join expert Janani Ravi to explore the fascinating world of natural language processing and Hugging Face. You’ll understand transformer models, the concept of attention, and how to leverage the powerful tools provided by Hugging Face. You'll gain hands-on experience with configuring tokenizers and fine-tuning models for specialized tasks like text classification. You’ll also dive into Hugging Face pipelines to perform complex tasks such as sentiment analysis, named entity recognition, text generation, and mask filling, all vital skills for today’s AI-driven industries.

What you’ll learn and how you can apply it

  • Learn the fundamentals of transformer models
  • Understand the structure of Hugging Face pipelines and use Hugging Face pipelines for NLP tasks
  • Fine-tune Hugging Face models to improve task-specific performance

This live event is for you because...

  • You’re familiar with machine learning and are curious about natural language processing with Hugging Face.
  • You’re an engineer who’s working with text-based models and you want to learn how to harness pretrained models for your use case.

Prerequisites

  • A Colab account for hands-on demos (free)
  • A Hugging Face account (free)
  • A level of comfort programming in Python
  • Experience building and training ML models
  • Some basic experience with natural language processing

Recommended follow-up:

Schedule

The time frames are only estimates and may vary according to how the class is progressing.

Introducing transformer models and Hugging Face (30 minutes)

  • Presentation: Quick overview of natural language processing; introducing transformers (encoder and decoder; attention layers; position encoding and embeddings); introducing Hugging Face (models, datasets, Spaces; the Inference API; creating an account and logging in)

Using Hugging Face pipelines for NLP (90 minutes)

  • Presentation: Structure of a Hugging Face pipeline; tokenizers, models, post-processing;
  • Demonstration: Sentiment analysis; name entity recognition; text summarization; text generation; mask filling
  • Group discussion: Other natural language models and use cases
  • Break

Tokenization and fine-tuning (60 minutes)

  • Presentation: Byte-pair encoding, WordPiece, and Unigram tokenizers
  • Demonstration: Configuring a BPE tokenizer; fine-tuning a text classification model; pushing the model to the Hugging Face Hub
  • Hands-on exercise: Explore user apps in Hugging Face Spaces

Your Instructor

  • Janani Ravi

    Janani Ravi is cofounder of Loonycorn, a team dedicated to upskilling IT professionals. She’s been involved in more than 100 online courses in data analytics, feature engineering, and machine learning. Previously, Janani worked at Google, Flipkart, and Microsoft. She completed her studies at Stanford.