TinyGBT(Tiny Gradient Boosted Trees) is a 200 line gradient boosted trees implementation written in pure python.
Since this code is not for production, it is not optimized for speed and memory usage.
- Data: LightGBM's regression example data
- TinyGBT is quite slower than LightGBM, but achieve almost same testset RMSE on similar parameter settings.
|RMSE of TestSet||0.45652||0.45934|
git clone https://github.com/lancifollia/tinygbt.git cd tinygbt python example.py
- run this code
- For now, Regression with L2 loss supported only.
-  T. Chen and C. Guestrin. XGBoost: A Scalable Tree Boosting System. 2016.
-  G. Ke et al. LightGBM: A Highly Efficient Gradient Boosting Decision Tree. 2017.