This repository collect some pretrained model for the implementation tf_chatbot_seq2seq_antilm
- example.tar.gz
(params: size = 128, vocab_size = 100000, num_layers = 4)
A pre-trained model using twitter_en corpus from this repository
File is splitted to walk-around file size limit, please combine them by:
cat example.tar.gz.part-* > example.tar.gz
- twitter_en_big_model.tar.gz
(params: size = 256, vocab_size = 100000, num_layers = 4)
The full model trained by twitter_en_big.txt.gz from this repository The network is bigger, and thus takes longer time to the predict.
File is splitted to walk-around file size limit, please combine them by:
cat twitter_en_big.txt.gz* > twitter_en_big.txt.gz