nlp-paper

Why we included this project

This repo is a reading list built from papers the author actually worked through, ordered by publication date so you can follow how a subfield changes over time. The standouts get detailed reading notes written in Chinese, cross-posted to Zhihu and CSDN, which help bridge the gap between the math in a paper and an intuitive understanding. Some classic models come with reproduction code in both TensorFlow and PyTorch, so you can go from reading to running without switching frameworks. A small search script keeps the archive navigable as it grows. It's a curated study companion plus reference implementations, not a drop-in production toolkit; the author keeps production-style code in separate sibling repositories.