Skip to content

Build a chatbot based on gender with Seq2Seq and Schedule-Sampling Seq2Seq with PyTorch

Notifications You must be signed in to change notification settings

purelyvivid/Schedule_Sampling_Seq2Seq_Movie_Script_Chatbot_PyTorch

Repository files navigation

Schedule Sampling Seq2Seq -- Movie_Script Chatbot -- PyTorch

Objective

A chatbot which can learn different speaking types between male/female from movie scripts

Dataset

Cornell movie-dialogue corpus

  • 220,579 conversational exchanges between 10,292 pairs of movie characters
  • involves 9,035 characters from 617 movies

Model

  1. Seq2Seq [1]

  1. Scheduled Sampling for Seq2Seq [2]

Result

  1. Loss

  1. Chatting...

Reference paper

  • [1] Sequence to Sequence Learning with Neural Networks, Ilya Sutskever, Oriol Vinyals, Quoc V. Le, 2014
  • [2] Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks, Samy Bengio, Oriol Vinyals, Navdeep Jaitly, Noam Shazeer, 2015

Releases

No releases published

Packages

No packages published