Hugging Face NLP Course transitions to LLM Course

Hugging Face is rebranding its long-standing NLP Course as the LLM Course to better reflect the shift toward Large Language Models and integrate advanced topics like reasoning models and fine-tuning. This update ensures the educational resources at hf.co/learn remain current with rapid AI breakthroughs while continuing to support fundamental NLP tasks.

Integration of LLM and Reasoning Model Research

The LLM Course expands beyond traditional Natural Language Processing to include state-of-the-art research and practical implementation. New chapters specifically address the development of reasoning models, such as Deepseek R1, and the fine-tuning of LLMs.

To ensure the material is accessible and community-driven, Hugging Face is moving toward a framework-agnostic approach. While the transformers library remains a primary reference, the course now integrates a variety of complementary tools and libraries, including:

  • LlamaIndex and LangChain for orchestration and retrieval.
  • Unsloth for optimized fine-tuning.
  • Marimo for interactive notebooks.
  • Contributions from community experts like Maxime Labonne.

Preservation of Classic NLP Foundations

Despite the shift toward LLMs, Hugging Face will maintain and modernize the existing material covering classic NLP tasks. The lab emphasizes that LLMs are not required for every application and that simpler, locally-run, and interpretable tasks remain highly valuable for students.

Upcoming updates to these classic chapters will include:

  • Sentence Transformers for embedding and similarity tasks.
  • ModernBert for efficient encoder-based architectures.
  • Updated approaches to Zero Shot classification.

Educational Format and Community Engagement

The course will prioritize reliable, long-term written material and coding exercises. However, Hugging Face will implement interactive exercises and host live sessions for high-engagement topics where synchronous learning is most effective.

Community members can contribute to the course by following the official organization on the Hugging Face Hub to suggest interactive units, propose live sessions, or join the development process.

Sources