llama-index-llms-huggingface 0.4.0

llama-index llms huggingface integration

LlamaIndex Llms Integration: Huggingface

Installation

1. Install the required Python packages:

   %pip install llama-index-llms-huggingface
   %pip install llama-index-llms-huggingface-api
   !pip install "transformers[torch]" "huggingface_hub[inference]"
   !pip install llama-index
   

2. Set the Hugging Face API token as an environment variable:

   export HUGGING_FACE_TOKEN=your_token_here
   

Usage

Import Required Libraries

import os
from typing import List, Optional
from llama_index.llms.huggingface import HuggingFaceLLM
from llama_index.llms.huggingface_api import HuggingFaceInferenceAPI

Run a Model Locally

To run the model locally on your machine:

locally_run = HuggingFaceLLM(model_name="HuggingFaceH4/zephyr-7b-alpha")

Run a Model Remotely

To run the model remotely using Hugging Face's Inference API:

HF_TOKEN: Optional[str] = os.getenv("HUGGING_FACE_TOKEN")
remotely_run = HuggingFaceInferenceAPI(
    model_name="HuggingFaceH4/zephyr-7b-alpha", token=HF_TOKEN
)

Anonymous Remote Execution

You can also use the Inference API anonymously without providing a token:

remotely_run_anon = HuggingFaceInferenceAPI(
    model_name="HuggingFaceH4/zephyr-7b-alpha"
)

Use Recommended Model

If you do not provide a model name, Hugging Face's recommended model is used:

remotely_run_recommended = HuggingFaceInferenceAPI(token=HF_TOKEN)

Generate Text Completion

To generate a text completion using the remote model:

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

Set Global Tokenizer

If you modify the LLM, ensure you change the global tokenizer to match:

from llama_index.core import set_global_tokenizer
from transformers import AutoTokenizer

set_global_tokenizer(
    AutoTokenizer.from_pretrained("HuggingFaceH4/zephyr-7b-alpha").encode
)

LLM Implementation example

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