Hi,
Is it possible to use a bi-directional RNN in the first layer, and uni-directional RNN in the remaining layers of the encoder as done by Wu et al. (2016): ?
Hi,
Is it possible to use a bi-directional RNN in the first layer, and uni-directional RNN in the remaining layers of the encoder as done by Wu et al. (2016): ?
Hi,
No, this is not currently implemented but it could be quite easily added in the future.
Available architectures are described here.
I just added the Google’s NMT encoder by the way. See:
http://opennmt.net/OpenNMT/training/models/#googles-nmt-encoder
You also noticed the new -encoder_type
option that replaces the boolean flags -brnn
, -dbrnn
and -pdbrnn
.