probtorch / pytorch

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Home Page:http://pytorch.org

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Implement Student's t-distribution

alicanb opened this issue · comments

@alicanb Do you it's worth implementing a standard_t_grad() function so that StudentT has deterministic reparameterized gradients? It took Martin and I about two solid weeks to implement each of standard_gamma_grad() and dirichlet_grad(), but at this point it would probably take less than a week to implement a standard_t_grad.

This can be closed. @fritzo @alicanb