tmbdev/clstm

Missing last char of the line.

ShubhamAgarwal12 opened this issue · 4 comments

If I train "lstm1" instead of "bidi", it is almost always missing the last character in the output sentence. Is it expected behaviour?

May I ask why do you train with 'lstm1'?

I need a network which can predict faster. Bidirectional net is very slow.

You can make clstm run >x2 faster!
#70 (comment)

Thanks Amit!