zyds/transformers-code
A practical course repository teaching Huggingface Transformers with Jupyter notebooks covering NLP tasks, model fine-tuning with PEFT, and distributed training.

A hands-on learning resource focused on the Huggingface Transformers library, synchronized with video tutorials on Bilibili and YouTube. The repository covers foundational NLP components including Pipeline, Tokenizer, and Model, and progresses through advanced topics like PEFT methods, low-precision training with bitsandbytes, and distributed training using accelerate. Exercises include text classification, named entity recognition, text summarization, and building dialogue systems.