Skip to content

Utkichaps/HAL-Mimic-2.0

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 

Repository files navigation

HAL-Mimic-2.0

This is an improvement over my existing Hal Mimic chatbot. This uses an Encoder-Decoder model along with an attention layer for better context responses compared to the previous chatbot which used two bi-directional LSTM layers.

You can check out the previous chatbot here

The model being used here is inspired by the Neural Machine Translation model used in the Natural Language Processing Specialization by Deep Learning.ai (Course 4 - Attention Models) and has been tweaked to be used as a chatbot

Point to remember: As this a more complex model than the previous one, for best results you will need a lot of data to train it (I tried it with around 3 years of chat data with a few friends I text). If you have a smaller dataset, you can try my previous chatbot. That works well with less data too.

About

This chatbot is an update over HAL-Mimic. It uses an encoder-decoder architecture with an attention layer and can be trained on your Whatsapp data to text like you.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors