Studies ML

Installation

1. Visit https://pytorch.org/ and select a suitable libtorch version to download.

   For example:

   # If you need e.g. CUDA 9.0 support, please replace "cpu" with "cu90" in the URL below.
   wget https://download.pytorch.org/libtorch/nightly/cpu/libtorch-shared-with-deps-latest.zip
   unzip libtorch-shared-with-deps-latest.zip

2. Place the unzipped libtorch folder under the root of the project directory，and pick either one below

   # list(APPEND CMAKE_PREFIX_PATH "libtorch")
   list(APPEND Torch_DIR, "libtorch")

   ahead of find_package(Torch REQUIRED) in CMakeLists.txt.

Note: For Mac M1 user, besides following the steps above, we'll have several extra steps to be done. Check this.

Contents

Note: most of the comments and .tex files are written in Chinese [ch-CN].

Notes

1. Studies of the Watermelon Book: Main entry & PDF

   • Model evaluation and selection

   • Linear model

   • Decision tree

   • Neural networks

2. Studies of Essential Math for AI: Main entry & PDF.

   • Data

   • Singular value decomposition

Codes

1. Starter

   • Pytorch C++ API

   • Using the PyTorch C++ Frontend

2. Basics

   • basics: cpp & py

   • linear regression: cpp & py

   • logistic regression: cpp & py

   • feed forward neural net: cpp & py

3. Intermediate

   • convolutional neural network: cpp & py

   • deep residual network: cpp & py

4. Advanced

5. Interactive

Project Structure

CMake deps

• extern

• utils

References

1. WatermelonBook [ch-CN]: https://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/MLbook2016.htm

2. PumpkinBook [ch-CN]: https://github.com/datawhalechina/pumpkin-book

3. Essential Math for AI [en-US]: https://www.oreilly.com/library/view/essential-math-for/9781098107628/

4. PyTorch: https://pytorch.org/

5. PyTorch Cpp and PyTorch Py