PrincetonVision / marvin

Marvin: A Minimalist GPU-only N-Dimensional ConvNets Framework
MIT License
421 stars 137 forks source link

Add AdaGrad (adaptive gradient algorithm) as an optimizer #6

Closed danielsuo closed 8 years ago

danielsuo commented 8 years ago

https://en.wikipedia.org/wiki/Stochastic_gradient_descent#AdaGrad http://www.magicbroom.info/Papers/DuchiHaSi10.pdf