This is a library that collects the useful codes for data science competition, as well as some code for research purposes.
The library contains materials for Tabular Data Mining, NLP, CV and RL.
Most of the code are written in Python 3.7, except some part of Cython, C++ and R. We do, however, provide an unified interface for python.
Some reference for what will be implemented can be found here https://www.overleaf.com/read/xftrzgtcxkpd (registration needed). I will be hosting the details on the lecture on GeekBang (https://time.geekbang.org/).
It is suggested if you want to use PyTorch and related libraries, to download the docker using the command docker pull opencompetition/opencompetition:0.1. Note that one should use NVidia-docker instead of docker.
To use the tensorflow libraries, it is recommended to use the official google cloud vm's support TPU training. Similarly, if one wishes to use PyTorch XLA to run PyTorch code on GPU, use Google Cloud VM as well.