MyMagic AI LLM

Introduction

This notebook demonstrates how to use MyMagicAI for batch inference on massive data stored in cloud buckets. The only enpoints implemented are complete and acomplete which can work on many use cases including Completion, Summariation and Extraction. To use this notebook, you need an API key (Personal Access Token) from MyMagicAI and data stored in cloud buckets. Sign up by clicking Get Started at MyMagicAI’s website to get your API key.

Setup

To set up your bucket and grant MyMagic API a secure access to your cloud storage, please visit MyMagic docs for reference. If you’re opening this Notebook on colab, you will probably need to install LlamaIndex 🦙.

%pip install llama-index-llms-mymagic

!pip install llama-index

from llama_index.llms.mymagic import MyMagicAI

llm = MyMagicAI(
    api_key="your-api-key",
    storage_provider="s3",  # s3, gcs
    bucket_name="your-bucket-name",
    session="your-session-name",  # files should be located in this folder on which batch inference will be run
    role_arn="your-role-arn",
    system_prompt="your-system-prompt",
    region="your-bucket-region",
)

resp = llm.complete(
    question="your-question",
    model="chhoose-model",  # currently we support mistral7b, llama7b, mixtral8x7b,codellama70b, llama70b, more to come...
    max_tokens=5,  # number of tokens to generate, default is 10
)

# The response indicated that the final output is stored in your bucket or raises an exception if the job failed
print(resp)

Asynchronous Requests by using acomplete endpoint

For asynchronous operations, use the following approach.

import asyncio

async def main():
    allm = MyMagicAI(
        api_key="your-api-key",
        storage_provider="s3",  # s3, gcs
        bucket_name="your-bucket-name",
        session="your-session-name",  # files should be located in this folder on which batch inference will be run
        role_arn="your-role-arn",
        system_prompt="your-system-prompt",
        region="your-bucket-region",
    )
    response = await allm.acomplete(
        question="your-question",
        model="chhoose-model",  # currently we support mistral7b, llama7b, mixtral8x7b,codellama70b, llama70b, more to come...
        max_tokens=5,  # number of tokens to generate, default is 10
    )

    print("Async completion response:", response)

await main()