Antonio Valerio Miceli Barone
Antonio Valerio Miceli Barone
In [domain_iterator.py](https://github.com/rsennrich/nematus/blob/master/nematus/data_iterator.py#L145l) and [domain_interpolation_data_iterator.py](https://github.com/rsennrich/nematus/blob/master/nematus/domain_interpolation_data_iterator.py#L183) there is this fragment of code: if len(ss) > self.maxlen and len(tt) > self.maxlen: continue which skips a sentence pair if **both** the source sentence and...
In [this branch](https://github.com/rsennrich/nematus/tree/floatX), I removed all hardcoded references to float32 and I tried to train with float16, but it does not work: Using cuDNN version 5105 on context None Mapped...
Shouldn't alpha be an expression with the same dimension as the input, rather than a float? https://github.com/marian-nmt/marian-dev/blob/2e77d5ecebcf48dd4114878cc9c136a7d09c9823/src/graph/expression_operators.cpp#L31