LLM Course
Hugging Face
Hugging Face’s course on its open-source libraries for language models. The first chapters cover Transformers, the Hub, Datasets and Tokenizers, and how to fine-tune a pretrained model. Later chapters move on to classic NLP tasks, model demos with Gradio, and advanced LLM topics such as dataset curation and reasoning models. Each chapter is planned as about a week of work, the notebooks run in Google Colab, and you should already know Python well.
What you'll learn
- Using pretrained models from the Hugging Face Hub with the pipeline API
- How encoder, decoder and encoder-decoder Transformers work
- Fine-tuning a model on your own dataset
- Processing data with the Datasets and Tokenizers libraries
- Building and sharing model demos
- Fine-tuning LLMs and building reasoning models
