ponytAIl: Polymorphic Orchestration of Networked Youthful, Adaptable Intelligence with Language.
Project description
ponytAIl
Then God blessed them and said, 窶廝e fruitful and multiply." - Genesis 1:28
ponytAIl: Polymorphic Orchestration of Networked Youthful, Adaptable Intelligence with Language. It is an node based multi-agent system targeting both LLM and BCI (in the future)
Installation
Install ponytAIl using pip:
pip install ponytail-agents
Usage
The ponytail-agents
command invokes the main
function with the following command-line arguments:
-f
,--file_path
(Required): Path to the input folder containing the necessary files (concluder.md
,starter.md
,node_creator.md
, and any additional nodes).-g
,--goal
(Required): Goal or objective of the task.-m
,--model
(Optional, default: "random-fast"): Name of the model to use. For available models, refer to the FLUTE repository.-r
,--result
(Optional): Additional result or output.
Example Commands
ponytail -f "YOUR_DIRECTORY\starter.md" -g "Define the number 1 using the mathematical collection"
ponytail -f "YOUR_DIRECTORY\starter.md" -g "Generate 5 candidate names for the self-reproducive LLM based multi-agent system. Note that the name must be abbreviated to 'PONYTAIL.'"
ponytail -f "YOUR_DIRECTORY\starter.md" -g "You are a villager. Send CREATE_NODES request by outlining other villagers, until you are satisfied with the diversity of the community. Use their names as the file names. Setup their profiles (ex. jobs, families, age, gender, etc) in a realistic and detailed manner. After, and ONLY after you are satisfied with the listed villagers, SEND_TO concluder.md to end the process."
Prerequisites
To run ponytAIl, prepare a folder containing all the files from nodes_sample
: concluder.md
, starter.md
, and node_creator.md
.
If you want to include additional nodes from the beginning, create them following the format of medium_sample.md
.
Growing a Node Community
By continuing to use the same folder after running the command once, you can nurture a node community over time. The system will build upon the existing nodes and generate new ones based on the interactions.
Interrupting the Process
If you need to interrupt the process at a reasonable point, as it may take a long time, use the following key combinations:
- Windows/Linux:
Ctrl+C
- Mac:
Cmd+C
Output
The output will be saved as full_results_YYYYMMDDHHmmss.md
in the parent folder of the starting folder.
LICENSE
The repository is licensed under the latest version of Modular and Inclusive Software Advancement License Classic (MISA-CLASSIC License).
There are 4 main policies that consist of this license.
- Disclaimer of Liability
- Naming Continuity Obligation
- Waiver of Other Copyrights
- Modular Extensibility (Defines how to modify the license)
See the license document for more details.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for ponytail_agents-0.1.6-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 2e1289e8f2e4175e8cf7f2de6530caf8361eeba4893b2631217ba4566ddb2845 |
|
MD5 | 3b560bd85aa5e38fab644924d9edd108 |
|
BLAKE2b-256 | 03f0d96ecb736ddeaf6c0f7ee4179825bffdbf49c8531aa57c8035119c3f7c10 |