Building AGI loops using LlamaIndex and Langchain
Project description
🤖 Llama AGI 🦙
This python package allows you to quickly create Auto-GPT-like agents, using LlamaIndex and Langchain.
Setup
Install using pip:
pip install llama-agi
Or install from source:
git clone https://github.com/run-llama/llama-lab.git
cd llama-lab/llama_agi
pip install -e .
Example Usage
The following shows an example of setting up the AutoAGIRunner
, which will continue completing tasks (nearly) indefinitely, trying to achieve it's initial objective of "Solve world hunger."
from langchain.agents import load_tools
from langchain.llms import OpenAI
from llama_agi.execution_agent import ToolExecutionAgent
from llama_agi.runners import AutoAGIRunner
from llama_agi.task_manager import LlamaTaskManager
from llama_agi.tools import search_notes, record_note, search_webpage
from llama_index import ServiceContext, LLMPredictor
# LLM setup
llm = OpenAI(temperature=0, model_name='text-davinci-003')
service_context = ServiceContext.from_defaults(llm_predictor=LLMPredictor(llm=llm), chunk_size_limit=512)
# llama_agi setup
task_manager = LlamaTaskManager([args.initial_task], task_service_context=service_context)
tools = load_tools(["google-search-results-json"])
tools = tools + [search_notes, record_note, search_webpage]
execution_agent = ToolExecutionAgent(llm=llm, tools=tools)
# launch the auto runner
runner = AutoAGIRunner(task_manager, execution_agent)
objective = "Solve world hunger"
initial_task = "Create a list of tasks"
sleep_time = 2
runner.run(objective, initial_task, sleep_time)
More examples can be found in the examples
folder!
Llama Ecosystem
- LlamaIndex (connecting your LLMs to data): https://github.com/jerryjliu/llama_index
- LlamaHub (community library of data loaders): https://llamahub.ai
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
llama_agi-0.2.0.tar.gz
(10.5 kB
view hashes)
Built Distribution
llama_agi-0.2.0-py3-none-any.whl
(17.1 kB
view hashes)
Close
Hashes for llama_agi-0.2.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 96f16734956414e56fea39f9cfadce8372c73ae23f9f819e251be7938b80ef9d |
|
MD5 | 8a77d4eeb4bf6692bc4a9201beed32d5 |
|
BLAKE2b-256 | 4440d3b611762c36ce27012df0ebdef1734d851f7169a2811fd505bcb484a8ae |