Skip to content

iivek/pop-art-deepmind

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Pop-Art paper by DeepMind

About

Reproducing the example from the paper [1] (tribute).

In addition, we compare Normalized SGD to Pop-Art SGD; while the former uses gradient rescaling and the latter is based on rescaling weights, the two are equivalent in case of squared loss.

pop-art

Run

To build the Docker image and run the example, use

make run

References

[1] Hasselt et al. Learning values across many orders of magnitude. NIPS 2016.

Releases

No releases published

Packages

No packages published