search

pskrunner14 / translator

PyTorch implementation of Machine Translation using Autoencoder with Attention
MIT License
9 stars 3 forks source link

Variable sized mini-batches #1

Open pskrunner14 opened 5 years ago

pskrunner14 commented 5 years ago

Model can't handle variable size mini-batches

pskrunner14 commented 5 years ago

Add torch.nn.utils.rnn.pad_sequence as preprocessing part and use packing and padding in encoder