// the find
susanli2016/NLP-with-Python
Scikit-Learn, NLTK, Spacy, Gensim, Textblob and more
A flat collection of Jupyter notebooks covering classical NLP techniques — NLTK, spaCy, Gensim, scikit-learn, and some early deep learning with Keras/TF 2.0. It reads like a study journal from a practitioner working through the NLP stack circa 2019–2021. Useful as a reference for someone learning the pre-LLM NLP toolchain.
Covers a genuinely wide range of techniques in one place: BoW/TF-IDF, LDA, Word2Vec, Doc2Vec, HMMs, NER, LSTM-based classifiers, and interpretability with LIME/SHAP. Real datasets (Amazon reviews, Quora, consumer complaints) make the examples more instructive than toy data. The BERT classification notebook is a practical baseline that still holds up for fine-tuning tasks. Breadth across libraries means you can compare approaches side-by-side without hunting across multiple repos.
The README is two lines — no index, no prerequisites, no explanation of what any notebook covers. You have to open each one to find out what it does. Most techniques are pre-transformer and the repo hasn't been meaningfully updated since 2021; nothing here on sentence transformers, LLM-based NLP, or modern embedding models. No requirements file or environment spec, so getting notebooks to run is an exercise in dependency archaeology — library APIs have shifted significantly since these were written. It's a personal notebook dump, not a structured learning path, so there's no progression or narrative connecting the material.