guxd / DialogWAE

Source Code for DialogWAE: Multimodal Response Generation with Conditional Wasserstein Autoencoder (https://arxiv.org/abs/1805.12352)

guxd/DialogWAE Issues

Code explanation about data prepocessing
Updated 4 years ago2
It is reasonable if I set n_samples to 1 when I run the sample.py
Updated 5 years ago2
Could I apply for your pretrained model in DailyDialog?
Closed 5 years ago
seems the loss of both generator and discriminator would collapse?
Updated 5 years ago1
Warning when run sample.py：RNN module weights are not part of single contiguous chunk of memory.
Closed 5 years ago3
Confused about the evaluation of inter-dist metrics.
Closed 5 years ago4
Wasserstein distance between prior and posterior
Updated 5 years ago3
Unable to achieve published result in DailyDialogue
Updated 4 years ago3
SWDA seems to give lot of repetition in the sample responses for the test data.
Updated 6 years ago4
Code explanation
Closed 6 years ago3
Stop training the context during train_G/train_D?
Closed 6 years ago3