Streaming chatbot

Changes to Anyscale Endpoints API

Effective August 1, 2024 Anyscale Endpoints API will be available exclusively through the fully Hosted Anyscale Platform. Multi-tenant access to LLM models will be removed.

With the Hosted Anyscale Platform, you can access the latest GPUs billed by the second, and deploy models on your own dedicated instances. Enjoy full customization to build your end-to-end applications with Anyscale. Get started today.

Add your Anyscale Endpoints token and run the following code to create a local chatbot.

import sys
import openai

ANYSCALE_ENDPOINT_TOKEN = "YOUR_ANYSCALE_TOKEN"
class OpenAIChatAgent:
    def __init__(self, model: str):
	#This simple example, doesn't modify the past conversation.
	#Eventually you run out of context window, but this should be enough for a 30-step conversation.
	#You need to either trim the message history or summarize it for longer conversations.
        self.message_history = []
        self.model = model
        self.oai_client = openai.OpenAI(
           base_url = "https://api.endpoints.anyscale.com/v1",
           api_key=ANYSCALE_ENDPOINT_TOKEN
        )
    def greet(self):
        return None

    def process_input(self, input: str):
        self.update_message_history(input)

        response = self.oai_client.chat.completions.create(

           model = self.model,
           messages = self.message_history,
           stream = True
        )
        words = ''
        for tok in response:
            delta = tok.choices[0].delta
            if not delta: # End token
                self.message_history.append({
                    'role': 'assistant',
                    'content': words
                })
                break
            elif delta.content:
                words += delta.content
                yield delta.content
            else:
                continue

    def update_message_history(self, inp):
        self.message_history.append({
            'role': 'user',
            'content': inp
        })

agent = OpenAIChatAgent("meta-llama/Llama-2-70b-chat-hf")
sys.stdout.write("Let's have a chat. (Enter `quit` to exit)\n")
while True:
    sys.stdout.write('> ')
    inp = input()
    if inp == 'quit':
        break
    for word in agent.process_input(inp):
        sys.stdout.write(word)
        sys.stdout.flush()
    sys.stdout.write('\n')