[!CAUTION] DISCLAIMER
All cheatsheets and documentation contained in this repository were generated by Google’s AI, “Antigravity”.
- AI Generated: The content is entirely AI-generated and may not have undergone rigorous human review.
- No Warranty: The author makes no warranty, express or implied, regarding the accuracy, completeness, safety, or reliability of the content.
- Use at Your Own Risk: The author assumes no liability for any damages, data loss, system failures, or other disadvantages arising from the use of the information in this repository. You use this information entirely at your own risk.
- Potential for Errors: The code and theoretical explanations may contain errors (hallucinations). Always verify with official documentation or expert sources.
This repository collects technical cheatsheets covering machine learning, numerical computing, and software engineering. Each file focuses on a specific topic, ranging from theoretical backgrounds to practical code examples.
| File | Topic | Summary |
|---|---|---|
| Flax NNX | Deep Learning | A guide specialized for NNX, the new API for Flax. It explicitly avoids the Linen API and explains model definition, state management, and training loops in a Pythonic object-oriented style. |
| Numerical Interpolation | Numerical Analysis | A summary of interpolation methods using scipy and JAX-based interpax. Covers theory and implementation comparisons for Lagrange interpolation, Spline interpolation (Cubic, PCHIP, Akima), and multidimensional interpolation. |
| JAX | High Perf. Computing | A comprehensive guide to JAX core concepts (jit, grad, vmap), random number generation, control flow (scan, cond), and distributed processing (Sharding). |
| Linear Solver | Linear Algebra | Comparison and implementation of linear algebra solvers in NumPy, JAX (Lineax), Julia, and PyTorch. Covers theoretical properties and usage of direct methods (LU, Cholesky, QR, TDMA) and iterative methods (CG, GMRES). |
| MLflow | MLOps | A “raw” usage guide for the MLflow Python client, avoiding auto-logging integrations. Detailed coverage of manual tracking, nested runs, and post-experiment analysis (programmatically finding/retrieving best runs). |
| Neural Network | Deep Learning Theory | From the basics of neural networks to the latest theories. Covers activation functions, ResNet (Gradient Highway), math of RNN/LSTM/Transformer, GNNs, Neural ODEs, and KAN (Kolmogorov-Arnold Networks). |
| Optax | Optimization / JAX | Gradient processing and optimization library for JAX. Covers core concepts, common optimizers, schedules, loss functions, and integration with Flax (Linen/NNX) and Muon. |
| Option Theory | Quantitative Finance | A summary of option pricing theory in financial engineering. Covers stochastic differential equations, Black-Scholes model, various exotic options (Barrier, Asian), and quantitative simulation of delta hedging. |
| Polars | Data Science | A complete guide to the high-performance DataFrame library Polars. Explains differences from Pandas, query writing with Expressions, optimization via Lazy API, and tips for performance. |
| Productivity & Philosophy | Soft Skills / Philosophy | A guide to engineering philosophy and productivity, covering Deep Work, complexity management, teamwork (HRT), problem solving (McKinsey style), and professional communication. |
| Quant Sins | Quantitative Finance | A summary of the “Seven Sins” of quantitative investing (backtesting biases) and their remedies, based on Deutsche Bank research. |
| Research Hacks | Research & Writing | Research productivity hacks (tools, workflow), guidelines for writing compelling Abstracts/Introductions, and inspiring quotes for scientists. |
| Ruff | Code Quality | A cheat sheet for Ruff (modern Python linter/formatter) with uv. Covers installation, pyproject.toml configuration, VSCode settings, and strict type hint/docstring enforcement. |
| Scikit-learn | Machine Learning | A comprehensive 33-part cheatsheet collection covering the entire scikit-learn ecosystem. Includes a high-level overview and detailed guides for all major modules. |
| Technical Writing | Communication | A style guide combining Google’s technical writing principles (Active Voice, BLUF) and Japanese technical writing (“理科系の作文技術”). Covers core principles, document structure, and specific rules for clear English and Japanese technical text. |
| uv | Python Tooling | Usage of uv, the fast Python package manager. Covers project initialization, dependency management, script execution, tool management, and Python version management. |
Generated by Google Antigravity