This is the code for paper "Generating Responses with a Specific Emotion in Dialog" in acl 2019.
This code is written in tensorflow, the start file is exp/train_batch.py. You can use the following commands to train this model:
python train_batch.py
--config_file configguration_file
--pre_train_word_count_file emotional_word_count_file_in_supervised_emotional_data_with_"word|||count"
--emotion_words_dir this_is_the_emotional_words_directory_with_six_emotional_word_files
--post_file post_file_name
--response_file response_file_name
--emotion_label_file emotion_label_file_for response
--embedding_file word_embedding_file
--train_word_count total_word_count_file_in_stc_data_with_"word|||count"
--checkpoint_dir saving_dir