Skip to content

Latest commit

 

History

History
77 lines (49 loc) · 2.69 KB

l8-chatbot.md

File metadata and controls

77 lines (49 loc) · 2.69 KB

Lesson8: Chatbot

Overview

In this notebook, you will explore how you can utilize the chat format to have extended conversations with chatbots personalized or specialized for specific tasks or behaviors.

OpenAI API call Architecture

  • Getting response from a single prompt/message :

  • Chat model

Setting multiple roles

In this lesson we're going to actually use a different helper function and instead of kind of putting a single prompt as input and getting a single completion, we're going to pass in a list of messages and these messages can be kind of from a variety of different roles

def get_completion_from_messages(messages, model="gpt-3.5-turbo", temperature=0):
    response = openai.ChatCompletion.create(
        model=model,
        messages=messages,
        temperature=temperature, # this is the degree of randomness of the model's output
    )
#     print(str(response.choices[0].message))
    return response.choices[0].message["content"]

  • The system message is a high-level instruction which sets the behavior of the assistant
  • The user message give the user instruction
  • The Assistant the model completion

Improvement interaction using context!

If you want the model to draw from, or quote, unquote remember earlier parts of a conversation, you must provide the earlier exchanges in the input to the model. And so we'll refer to this as context

messages =  [  
{'role':'system', 'content':'You are friendly chatbot.'},
{'role':'user', 'content':'Hi, my name is Isa'},
{'role':'assistant', 'content': "Hi Isa! It's nice to meet you. \
Is there anything I can help you with today?"},
{'role':'user', 'content':'Yes, you can remind me, What is my name?'}  ]
response = get_completion_from_messages(messages, temperature=1)
print(response)

Completion : 

Your name is Isa.

Building a OrderBot for a restaurant

We can automate the collection of user prompts and assistant responses to build a OrderBot. The OrderBot will take orders at a pizza restaurant.

  • Check the notebook to see the entire implementation.

References

Main course :

Transformer paper - Attention is all you need: