npinto / asgd

Averaged Stochastic Gradient Descent Classifiers
41 stars 21 forks source link

Use η0 / (1 + λ η0 t)^0.75 by default instead of ...^2/3. #12

Open npinto opened 13 years ago

npinto commented 13 years ago

The learning rate has the form η0 / (1 + λ η0 t)^0.75 where λ is the regularization constant.

See: http://leon.bottou.org/projects/sgd