LLM Agent Composer
This tool allows developers to compose multiple Large Language Model agents into a single, cohesive system. It provides a flexible framework for combining the strengths of different LLM agents, enabling the creation of more accurate and robust models. The tool is useful for AI developers who want to leverage the benefits of ensemble learning and improve the overall performance of their LLM agents.
Installation
To install the required packages, run the following command:
pip install transformers scikit-learnUsage
To use the LLM Agent Composer, run the following command:
python llm_agent_composer.py --agents model1.pth model2.pth --composition_config composition_config.jsonReplace model1.pth and model2.pth with the paths to your LLM agent models, and composition_config.json with the path to your composition config file.
Source Code
import argparse
import json
from transformers import AutoModelForSequenceClassification
from sklearn.ensemble import VotingClassifier
from sklearn.linear_model import LogisticRegression
def compose_agents(agent_models, composition_config):
estimators = []
for model in agent_models:
estimator = (model, AutoModelForSequenceClassification.from_pretrained(model))
estimators.append(estimator)
if not agent_models:
return None
if 'weights' not in composition_config:
raise ValueError('Composition config must contain weights')
if len(agent_models) != len(composition_config['weights']):
raise ValueError('Number of agent models must match number of weights')
voting_classifier = VotingClassifier(estimators=estimators, voting='soft', weights=composition_config['weights'])
return voting_classifier
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='LLM Agent Composer')
parser.add_argument('--agents', nargs='+', help='List of LLM agent models')
parser.add_argument('--composition_config', help='Composition config file')
args = parser.parse_args()
with open(args.composition_config, 'r') as f:
composition_config = json.load(f)
composed_agent = compose_agents(args.agents, composition_config)
print(composed_agent)README
LLM Agent Composer
This tool allows developers to compose multiple Large Language Model agents into a single, cohesive system.
Installation
To install the required packages, run the following command:
pip install transformers scikit-learnUsage
To use the LLM Agent Composer, run the following command:
python llm_agent_composer.py --agents model1.pth model2.pth --composition_config composition_config.jsonReplace model1.pth and model2.pth with the paths to your LLM agent models, and composition_config.json with the path to your composition config file.
Composition Config File
The composition config file should be a JSON file with the following structure:
{
"weights": [0.5, 0.5]
}Replace [0.5, 0.5] with the weights for your LLM agent models.
Community
Downloads
···
Rate this tool
No ratings yet — be the first!
Details
- Tool Name
- llm_agent_composer
- Category
- LLM Agents
- Generated
- August 11, 2026
- Tests
- Passing ✅
- Fix Loops
- 5
Quick Install
Clone just this tool:
git clone --depth 1 --filter=blob:none --sparse \ https://github.com/ptulin/autoaiforge.git cd autoaiforge git sparse-checkout set generated_tools/2026-08-11/llm_agent_composer cd generated_tools/2026-08-11/llm_agent_composer pip install -r requirements.txt 2>/dev/null || true python llm_agent_composer.py