search

macournoyer / neuralconvo

Neural conversational model in Torch
776 stars 347 forks source link

Learning rate decay #47

Closed danpol closed 8 years ago

danpol commented 8 years ago

If optimState is reinitialized every epoch, LR will remain constant.

macournoyer commented 8 years ago

Thanks again!