llama-index-llms-openllm 0.4.0

llama-index llms openllm integration

LlamaIndex LLM Integration: OpenLLM

Installation

To install the required packages, run:

%pip install llama-index-llms-openllm
!pip install llama-index

Setup

Initialize OpenLLM

First, import the necessary libraries and set up your OpenLLM instance. Replace my-model, https://hostname.com/v1, and na with your model name, API base URL, and API key, respectively:

import os
from typing import List, Optional
from llama_index.llms.openllm import OpenLLM
from llama_index.core.llms import ChatMessage

llm = OpenLLM(
    model="my-model", api_base="https://hostname.com/v1", api_key="na"
)

Generate Completions

To generate a completion, use the complete method:

completion_response = llm.complete("To infinity, and")
print(completion_response)

Stream Completions

You can also stream completions using the stream_complete method:

async for it in llm.stream_complete(
    "The meaning of time is", max_new_tokens=128
):
    print(it, end="", flush=True)

Chat Functionality

OpenLLM supports chat APIs, allowing you to handle conversation-like interactions. Here’s how to use it:

Synchronous Chat

You can perform a synchronous chat by constructing a list of ChatMessage instances:

from llama_index.core.llms import ChatMessage

chat_messages = [
    ChatMessage(role="system", content="You are acting as Ernest Hemmingway."),
    ChatMessage(role="user", content="Hi there!"),
    ChatMessage(role="assistant", content="Yes?"),
    ChatMessage(role="user", content="What is the meaning of life?"),
]

for it in llm.chat(chat_messages):
    print(it.message.content, flush=True, end="")

Asynchronous Chat

To perform an asynchronous chat, use the astream_chat method:

async for it in llm.astream_chat(chat_messages):
    print(it.message.content, flush=True, end="")

LLM Implementation example

https://docs.llamaindex.ai/en/stable/examples/llm/openllm/