Llama Packs 🦙📦#
Concept#
Llama Packs are a community-driven hub of prepackaged modules/templates you can use to kickstart your LLM app.
This directly tackles a big pain point in building LLM apps; every use case requires cobbling together custom components and a lot of tuning/dev time. Our goal is to accelerate that through a community led effort.
They can be used in two ways:
On one hand, they are prepackaged modules that can be initialized with parameters and run out of the box to achieve a given use case (whether that’s a full RAG pipeline, application template, or more). You can also import submodules (e.g. LLMs, query engines) to use directly.
On the other hand, LlamaPacks are templates that you can inspect, modify, and use.
All packs are found on LlamaHub. Go to the dropdown menu and select “LlamaPacks” to filter by packs.
Please check the README of each pack for details on how to use. Example pack here.
See our launch blog post for more details.
Usage Pattern#
You can use Llama Packs through either the CLI or Python.
CLI:
llamaindex-cli download-llamapack <pack_name> --download-dir <pack_directory>
Python:
from llama_index.llama_pack import download_llama_pack
# download and install dependencies
pack_cls = download_llama_pack("<pack_name>", "<pack_directory>")
You can use the pack in different ways, either to inspect modules, run it e2e, or customize the templates.
# every pack is initialized with different args
pack = pack_cls(*args, **kwargs)
# get modules
modules = pack.get_modules()
display(modules)
# run (every pack will have different args)
output = pack.run(*args, **kwargs)
Importantly, you can/should also go into pack_directory to inspect the source files/customize it. That’s part of the point!