transformers-code

Why we included this project

This repo is the code companion to a Chinese-language video course on the Huggingface Transformers ecosystem. It starts with the basics: pipelines, tokenizers, models, datasets, and the Trainer API, then moves through concrete NLP tasks like named entity recognition, question answering, summarization, and retrieval-style chatbots. Later sections dig into the parts that tend to trip people up in practice: parameter-efficient fine-tuning with PEFT, low-precision and QLoRA training, and distributed training with Accelerate and DeepSpeed. The pinned dependency versions in the README mean you can run the examples without guesswork, and each lesson has a video link alongside the code. It is a learning path, not a production library, so expect guided practice rather than a drop-in component.