The genai from solonce

genai

A Command Line AI Assistant powered by Mixtral-8x7B-Instruct-v0.1 using the HuggingFace Inference API.

About the Project
Built With
Getting Started
- Prerequisites
- Installation
Flags
Examples
Limitations
Contributing
License
Authors

About The Project

This project is designed for fast and extremely lightweight AI inference when using a linux based machine right from the command line.

Using the command 'genai', you will be brought into a shell that allows you to interact with an AI Assistant. It is designed to give as short an answer as possible while satisfying the question.

Cool Features!

Shell support: Open a shell to pass multiple prompts to genai.
No shell/One-liner support: Use the -t flag to ask genai a single prompt without entering a shell
Context Operation: Genai runs under a given context so it knows how to format the response. Change it for a single shell instance or reset the default context to something custom. The default context is here.
Fast & Custom: Genai defaults to a lightweight 100 character response, with a maximum 3 second load time. These are variable with the correct flags if you need longer responses!

Built With

HuggingFace Inference API - Allow for free AI Inference

Mixtral-7B-Instruct-v0.2 - The Model

Getting Started

Getting started is very simple. All you need is to install 'genai-cmd' via pip, and supply a HuggingFace API Key.

Prerequisites

You will need a HuggingFace API Key. HuggingFace API Keys are free, and the instructions for obtaining one can be found here..

Installation

Get a free HuggingFace API Key here.
Use pip to install genai

pip install genai-cmd

Set the HuggingFace API Key

genai -S 'HF_API_KEY'

Have Fun! Use the following command to enter a shell!

genai

Flags

-t, --text-data      Used to ask a question to genai directly without entering the shell.
-r, --raw-data       Same as -t without text formatting.

-c, --context        Used to set the temporary context genai answers prompt(s) under. Works with shell. (ex. "You are a pirate", "You are an angry old man")
-l, --length         Used to set the maximum length of response. Default is 100 characters. Works with shell.
-s, --seconds        Used to set the maximum time a response will take to generate. Default is 3. Works with shell.

-C, --reset-context  Used to set the default context genai answers all prompts under. (ex. "You are a pirate", "You are an angry old man")
-S, --set-key        Used to set the HuggingFace Inference API key needed for genai to operate.

Examples

Entering a shell

The main usage of genai is to enter a shell that you can communicate with an AI Agent.

>genai
                                                     
                                                 88  
                                                 ""  
                                                     
    ,adPPYb,d8  ,adPPYba, 8b,dPPYba,  ,adPPYYba, 88  
   a8"    `Y88 a8P_____88 88P'   `"8a ""     `Y8 88  
   8b       88 8PP""""""" 88       88 ,adPPPPP88 88  
   "8a,   ,d88 "8b,   ,aa 88       88 88,    ,88 88  
    `"YbbdP"Y8  `"Ybbd8"' 88       88 `"8bbdP"Y8 88  
    aa,    ,88                                       
     "Y8bbdP"                                        

  ┌───────────────────────────────────────────────────────
  │ Version: 1.0.8
  │ Github: https://github.com/Solonce/genai
  │ Author: Solomon Ince
  │ 
  │ Model: https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2
  └───────────────────────────────────────────────────────
genai> How do I get only even numbers from an array in python?

  ┌─────────────────────────────
  │ [1, 2, 3, 4, 5]
  │ 
  │ filter(lambda x: x % 2 == 0, [1, 2, 3, 4, 5])
  └─────────────────────────────

genai> exit

Entering a shell with custom --length and --seconds

This is to demonstrate you can make large AI inference requests, even though this was built as a lightweight tool.

>genai -l 500 -s 15
                                                     
                                                 88  
                                                 ""  
                                                     
    ,adPPYb,d8  ,adPPYba, 8b,dPPYba,  ,adPPYYba, 88  
   a8"    `Y88 a8P_____88 88P'   `"8a ""     `Y8 88  
   8b       88 8PP""""""" 88       88 ,adPPPPP88 88  
   "8a,   ,d88 "8b,   ,aa 88       88 88,    ,88 88  
    `"YbbdP"Y8  `"Ybbd8"' 88       88 `"8bbdP"Y8 88  
    aa,    ,88                                       
     "Y8bbdP"                                        

  ┌───────────────────────────────────────────────────────
  │ Version: 1.0.8
  │ Github: https://github.com/Solonce/genai
  │ Author: Solomon Ince
  │ 
  │ Model: https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2
  └───────────────────────────────────────────────────────
genai> Write me a short paper on Artificial Society and its future.

  ┌─────────────────────────────
  │ I. Introduction
  │         Artificial Society (AS) refers to a virtual world populated by artificial
  │ agents, each with its own set of rules, behaviors, and interactions. AS is an
  │ extension of Artificial Intelligence (AI) and offers a platform for exploring complex
  │ social phenomena, testing theories, and simulating future scenarios.
  │ 
  │     II. Current State of Artificial Society
  │         AS is still in its infancy, with most research focusing on individual agents
  │ and their interactions. However, recent advancements in multi-agent systems, machine
  │ learning, and natural language processing have led to more sophisticated simulations
  │ of social structures and dynamics.
  │ 
  │     III. Potential Applications of Artificial Society
  │         1. Social Science Research: AS can be used to test theories and hypotheses in
  │ a controlled environment, providing valuable insights into human behavior and social
  │ dynamics.
  │         2. Education and Training: AS can be used to simulate real-world scenarios and
  │ provide students with a safe space to practice and learn social skills.
  │         3. Urban Planning and Design: AS can be used to simulate and optimize urban
  │ environments, taking into account factors such as traffic flow, resource allocation,
  │ and social interaction.
  │         4. Conflict Resolution and Negotiation: AS can be used to model and analyze
  │ complex negotiations and conflicts, providing insights into effective strategies and
  │ solutions.
  │ 
  │     IV. Future Directions for Artificial Society
  │         1. Scalability: Developing AS that can handle large-scale simulations with
  │ millions of agents and complex social structures.
  │         2. Realism: Incorporating more realistic models of human behavior and social
  │ dynamics, including emotions, motivations, and cultural norms.
  │         3. Ethics: Addressing ethical concerns related to AS, such as privacy,
  │ consent, and the potential for unintended consequences.
  │         4. Integration with Physical World: Exploring ways to integrate AS with the
  │ physical world, such as through augmented reality or the Internet of Things.
  │ 
  │     V. Conclusion
  │         Artificial Society offers a unique platform for exploring complex social
  │ phenomena and testing theories in a controlled environment. With continued research
  │ and development, AS has the potential to revolutionize fields such as social science,
  │ education, urban planning, and conflict resolution.
  └─────────────────────────────

genai> exit

One line inference

This is for fast inference if you just want one prompt answered.

>genai -t "How do I reverse an array in python"

   genai:
 ┌─────────────────────────────
 │ array = [1, 2, 3, 4, 5]
 │ array.reverse()
 └─────────────────────────────

One line inference with NO text formatting

This is the same, but instead of -t, we use -r for the raw response with no text formatting.

>genai -r "How do I reverse an array in python"

array = [1, 2, 3, 4, 5]
array.reverse()

One liner inference with temporary context included

This is one line text inference, but supplying context for the agent.

>genai --text-data "How was your day?" --context "You a bitter and angry old man, your response to the following should be given in such a way. "

   genai:
 ┌─────────────────────────────
 │ My day was a complete and utter waste of time, just like every other day in this miserable existence.
 │ I woke up this morning to the sound of my alarm clock, which is nothing more than a cruel reminder that I am
 │ still alive and have to endure another day of this meaningless life. I dragged myself out of bed and went through
 │ the motions of getting ready for work, all the while cursing the fact that I have to spend eight hours of my day in a cr
 └─────────────────────────────

Limitations

This project was never intended for a 'catch-all' or to be used as a perfect AI assistant. The goal was to simply create a command line AI assistant tool.
The Mixtral-7B-Instruct-v0.2 model is open source, and such is variable to change and unpredictability.
The Mixtral-7B-Instruct-v0.2 model tends to have an issue generating one response. It will often respond back as a user to prompt the model again.
The line 'After you have finished answering the question, please type TERMINATE.' is used in the default prompt, but not enforced with the -c or -C flags. This line can be important if you want nice text formatting.

Contributing

Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.

If you have suggestions for adding or removing projects, feel free to open an issue to discuss it, or directly create a pull request after you edit the README.md file with necessary changes.
Please make sure you check your spelling and grammar.
Create individual PR for each suggestion.
Please also read through the Code Of Conduct before posting your first idea as well.

Creating A Pull Request

Fork the Project
Create your Feature Branch (git checkout -b feature/AmazingFeature)
Commit your Changes (git commit -m 'Add some AmazingFeature')
Push to the Branch (git push origin feature/AmazingFeature)
Open a Pull Request

License

Distributed under the MIT License. See LICENSE for more information.

Authors

Solomon Ince - AI/ML Programmer - Solomon Ince - genai

solonce / genai Goto Github PK

genai's Introduction