Skip to content

Latest commit

 

History

History
115 lines (88 loc) · 3.95 KB

README.md

File metadata and controls

115 lines (88 loc) · 3.95 KB

Rosetta Stone Chatbot

GPT2-Telegram-Chatbot

GPT2 Telegram chatbot that's been relatively tuned for chatting. It could probably pass a turing test easily and I'd love for to get it turing test certified eventually. Feel free to make me PRs and I'll check out your code!

For a demo if it's online: https://telegram.me/finishsentencebot

Installation/How to use:

  1. Install gpt-2 as normal IE:

python3 -m pip install tensorflow

or

python3 -m pip install tensorflow-gpu

...

python3 -m pip install -r requirements.txt

  1. Place this file in the src/ folder.

  2. Install python-telegram-bot with

python3 -m pip install python-telegram-bot

  1. Input your telegram botfather bot key in the python file/

  2. run: python3 src/GPT2-Learning.py

If needed create a shell script that loops the bot python3 command in case of crashes. The bot requires atleast 8GB of free memory, (16GB preferred and a dual-core processor.)

For quick command reference:

start - Basic start command.
help - Show quick help command.
chatbot - Chatbot mode, me+you string edit and open-ended questions.
finish - Default finish sentence command counting words as output length.
learnon - Turn on sentence compilation with Me+you string for conversation.
learnoff - Turn off the sentnce compilation.
learnreset - Reset the learned conversation.

Example of learning mode backend:

INFO:tensorflow:Restoring parameters from models/774M/model.ckpt
2019-10-16 20:05:55,273 - tensorflow - INFO - Restoring parameters from models/774M/model.ckpt
==========
Before splitlines:  I'm in the middle of an internship! I'm a new intern here at Wh
==========
==========
Mode: True
Learn: True
Length: 17
==========
Before regex:  I'm in the middle of an internship! I'm a new intern here at Wh
==========
Output:  I'm in the middle of an internship!
==========
Raw_text or Original: Me: Hi there! My name is Pikachu, how are you today? You: Hello?  Me: Heya! What's up? You:
==========
Learning text or Next: Me: Hi there! My name is Pikachu, how are you today? You: Hello?  Me: Heya! What's up? You: I'm in the middle of an internship!
==========
INFO:tensorflow:Restoring parameters from models/774M/model.ckpt
2019-10-16 20:08:00,910 - tensorflow - INFO - Restoring parameters from models/774M/model.ckpt
==========
Before splitlines:  Interning is a great way to learn about a company and make some great connections!
==========
==========
Mode: True
Learn: True
Length: 17
==========
Before regex:  Interning is a great way to learn about a company and make some great connections!
==========
Output:  Interning is a great way to learn about a company and make some great connections!
==========
Raw_text or Original: Me: Hi there! My name is Pikachu, how are you today? You: Hello?  Me: Heya! What's up? You: I'm in the middle of an internship! Me: Oh that sounds fun! What's the internship about? You:
==========
Learning text or Next: Me: Hi there! My name is Pikachu, how are you today? You: Hello?  Me: Heya! What's up? You: I'm in the middle of an internship! Me: Oh that sounds fun! What's the internship about? You: Interning is a great way to learn about a company and make some great connections!
==========

For a list of grammarly scores please see /SCORES.MD.

Here's a example without the learning mode, that will need more work in the future:

You're a smelly tiger plush! I need to give you a bath.
Computing...
I'm not a smelly tiger plush.
Reply:
Yes you are! Hurry up and get in the bathtub, the water is warm!
Computing...
I'm not going to the bathtub, I'm going to the kitchen.
Reply:
Hey! You better get back here!
Computing...
I'm not going anywhere!
Reply:
Then get in the bath tub!
Computing...
I'm not going to get in the bath.```

Inspired by the book When HARLIE Was One by David Gerrold

HitCount