harvardnlp / seq2seq-attn

Sequence-to-sequence model with LSTM encoder/decoders and attention
http://nlp.seas.harvard.edu/code
MIT License
1.26k stars 278 forks source link

beam.lua should work with cuda or non-cuda #12

Closed srush closed 8 years ago

srush commented 8 years ago

Currently fails when reading a cuda tensor based model. Short-term it should probably "require 'cunn'", long-term we probably should support both modes.

nicolas-ivanov commented 8 years ago

@yoonkim any plans to reconsider this issue? As far as I know, beam.lua only works with cpu now.