bert-PostKS

bert-PostKS (Posterior Knowledge Selection)

Pytorch BERT augmented implementation of Learning to Select Knowledge for Response Generation in Dialog Systems

I replace original utterance encoder and knowledge encoder with pretrained BERT. For decoder, I apply Hierarchical Gated Fusion Unit (HGFU) [Yao et al. 2017] and I only use three number of knowledges for the sake of code simplicity.

Requirement

• pytorch
• pytorch_pretrained_bert

Train model

If you run train, vocab.json and trained parameters will be saved. Then you can play demo.

$ python train.py -pre_epoch 5 -n_epoch 15 -n_batch 32

Play demo

$ python demo.py

You need to type three knowledges and utterance. Then bot will reply!

# example
Type first Knowledge: i'm very athletic.
Type second Knowledge: i wear contacts.
Type third Knowledge: i have brown hair.

you: hi ! i work as a gourmet cook .
bot(response): i don't like carrots . i throw them away . # reponse can change based on training.

• If you type "change knowledge" at (you), you can retype three knowledges.
• If you type "exit" at (you), you can terminate demo.

DataSet

I only use "self_original_no_cands" in Persona-chat released by ParlAI