# Persian AI Model Manager Documentation

Welcome to the official documentation for the **Persian AI Model Manager**. This guide explains in detail the algorithms, workflow, and commands used to create and use AI models with Persian AI. The system offers both a browser-based interface and a terminal-based interface, providing flexibility and control over model training and deployment.

---

## Table of Contents

- [Overview](#overview)
- [Creating Models](#creating-models)
  - [Using the Browser Interface (`learnweb`)](#using-the-browser-interface-learnweb)
  - [Using the Terminal Interface (`learnchat`)](#using-the-terminal-interface-learnchat)
- [Using Models](#using-models)
  - [Listing Models](#listing-models)
  - [Running Model in Browser (`runweb`)](#running-model-in-browser-runweb)
  - [Running Model in Terminal (`runchat`)](#running-model-in-terminal-runchat)
- [Algorithmic Workflow](#algorithmic-workflow)
- [Conclusion](#conclusion)

---

## Overview

The Persian AI Model Manager provides a powerful, yet user-friendly environment for:
- **Creating New Models:** Train models interactively via a browser or directly from the terminal.
- **Using Models:** Deploy and interact with your trained models either in a browser-based chat interface or directly via the terminal.

This documentation covers all commands and workflows required to fully leverage the system.

---

## Creating Models

### Using the Browser Interface (`persianai learnweb`)

- **Command:**  
  ```bash
  persianai learnweb

• PersianAI Website: You can online chat with PersianAi 1.0.0 or download Persian Model Creator and see more info in persianaichat.ir

• Description:
  Launches an interactive web-based interface that allows you to create a new AI model directly in your browser. This mode is ideal for users who prefer a graphical user interface for setting up and monitoring the training process.

• Workflow:

  1. User Initiation: Execute persianai learnweb from the terminal.
  2. Browser Launch: The system automatically opens a browser window with the model creation interface.
  3. Data Input: Fill in the form with model details (e.g., model name, parameters, dataset selection).
  4. Training Process: The interface displays real-time progress of the training algorithm.
  5. Model Registration: Once training is complete, the new model is saved and added to your model list.

---

Using the Terminal Interface (persianai learnchat)

• Basic Command:

  persianai learnchat MODELNAME INFORMATION_FOLDER_PATH

• Advanced Command (for more control):

  persianai learnchat MODELNAME INFORMATION_FOLDER_PATH accuracy_param test_size_param number_test

• Description:
  Creates a new model using data from the specified folder. The basic command offers a quick way to initiate training in the terminal, while the advanced command lets you fine-tune parameters like accuracy, test size, and the number of tests for better control over the model’s performance.

• Workflow:

  1. Command Execution: Enter the appropriate command with the required parameters in the terminal.
  2. Data Loading: The system reads and processes data from the provided folder path.
  3. Parameter Configuration: Training parameters (such as desired accuracy, test size, etc.) are configured based on user input.
  4. Model Training: The terminal displays training progress via logs and status messages.
  5. Model Saving: Upon completion, the trained model is saved and registered in your model registry.

---

Using Models

Listing Models

• Command:

  persianai list

• Description:
  Displays a list of all the models you have created along with details such as model name, creation date, and training parameters.

---

Running Model in Browser (persianai runweb)

• Command:

  persianai runweb MODELNAME

• Description:
  Launches a browser-based chat interface for the specified model. This is particularly useful for interactive testing, demonstration, and real-time interaction with your model.

• Workflow:

  1. Model Retrieval: The system retrieves the specified model.
  2. Interface Launch: A browser window opens with the chat interface.
  3. Interaction: Send and receive messages interactively via the web interface.

---

Running Model in Terminal (persianai runchat)

• Command:

  persianai runchat MODELNAME

• Description:
  Runs the specified model directly in the terminal. This mode is designed for quick testing, debugging, and integration with other terminal-based workflows.

• Workflow:

  1. Model Retrieval: The system loads the model from the stored registry.
  2. Terminal Chat: The terminal is set up for a chat-like interaction, displaying incoming messages and accepting user input.

---

Algorithmic Workflow

Below is a high-level overview of the algorithmic steps involved in both creating and using models:

1. Model Creation:

   • Initialization: User initiates the command (either learnweb or learnchat).
   • Data Processing: The system reads the dataset from a specified folder.
   • Parameter Setup: User-defined parameters (or defaults) are set.
   • Training: The model is trained using machine learning techniques. In the browser mode, this process is visualized with progress indicators; in the terminal mode, logs are printed.
   • Registration: Once training completes, the model is saved and registered in the model list for future use.

2. Model Deployment and Interaction:

   • Model Selection: The user lists available models using persianai list.
   • Interface Launch: The model is launched either in a web-based chat interface (runweb) or a terminal-based chat interface (runchat).
   • Interaction Loop: The user sends messages to the model, and the model processes the input, generates a response using its trained parameters, and returns the answer.
   • Logging & History: Conversation history is maintained locally (in the browser or terminal) for the duration of the session.

---

Conclusion

The Persian AI Model Manager offers a comprehensive, flexible platform for creating and deploying AI models. Whether you prefer an interactive browser interface or the control of the terminal, the system provides robust tools to:

• Create new models with customizable parameters.
• Deploy models in both web and terminal environments.
• Interact with your models in real-time, with local history management.

For additional details, refer to our GitHub repository or contact the developer, Mohammad Taha Gorji.

---

Happy Modeling!

Coded By Mohammad Taha Gorji