-
-
v0.2 Release: v0.2
Released code supports both Tensorflow and Pytorch (seq2seq and attention-seq2seq). Inference time of : * pytorch-gru: 0.4s, * pytorch-lstm: 0.7s, * pytorch-attention-gru: 1.2s.
-
v0.1 Release: v0.1
Support both Tensorflow and Pytorch, jointly train, and separately save encoder and decoder.