search

microsoft / EdgeML

This repository provides code for machine learning algorithms for edge devices developed at Microsoft Research India.
Other
1.59k stars 370 forks source link

Added Heirarchical prior. #247

Open meetdoshi90 opened 3 years ago

meetdoshi90 commented 3 years ago

Added path smoothing to prevent overfitting for deeper trees whilst improving accuracy. Added mse loss for regression on weighted layers. References:

  1. Path smoothing
  2. IHT Gradient descent
ghost commented 3 years ago

CLA assistant check
All CLA requirements met.

adityakusupati commented 3 years ago

hi @meetdoshi90,

This is very useful and seems correct. Thanks for the contribution. However, I think the TF bonsai version is obsolete due to its old dependencies. I think more people use the pytorch version more often.

@harsha-simhadri you might want to take a look at this when you get a chance.

Aditya

meetdoshi90 commented 3 years ago

Hey @adityakusupati I will implement the same on pytorch. Thanks for the advice. Meet.