21 1 year ago

This AI model is designed to automate and optimize code generation for industrial applications, helping to streamline manufacturing processes, enhance automation, and address complex problem-solving scenarios on the factory floor.

ollama run chrisdauksza/chris_dauksza_ai

Models

View all →

Readme

# Chris Dauksza AI Model

This repository contains a custom AI model tailored for generating complete, functional, and optimized code, designed with 32,768-token context and robust configuration options. The model reflects the expertise and communication style of Chris Dauksza, the founder of StarMaint AI, StarReliability AI, and EZMaint.

## Overview

The Chris Dauksza AI model is based on the "huihui_ai/qwq-fusion" model and is optimized for:
- Generating functional and complete code.
- Handling large contexts for multi-step problem-solving.
- Precision in responses tailored to industrial AI and automation applications.

## Features

- **32,768-Token Context**: Supports extensive context windows for handling complex tasks and long prompts.
- **Code Optimization**: Fine-tuned to produce clean, efficient, and logically structured code.
- **Dynamic Configuration**: Includes options for controlling temperature, sampling, and repetition penalties.
- **Multimodal Support**: Accepts text and image inputs for advanced use cases.
- **Tool Integration**: Enables calling external tools and functions for extended capabilities.

## Model Specifications

- **Base Model**: `huihui_ai/qwq-fusion`
- **Token Limit**: 32,768
- **Temperature**: 0.2 (optimized for deterministic outputs)
- **Sampling Methods**:
  - `top_k`: 20
  - `top_p`: 0.6
  - `repeat_penalty`: 1.2
- **Custom System Prompt**: Encodes the expertise and communication style of Chris Dauksza.

## Usage

### Prerequisites

- Install [Ollama CLI](https://ollama.com/docs/cli) for managing and running models.

### Creating the Model

1. Save the `Modelfile` configuration (provided in this repository) to a file named `Modelfile`.
2. Run the following command:
   ```bash
   ollama create chris_dauksza_ai -f ./Modelfile

Running the Model

Once the model is created, you can interact with it using the following command:

ollama run chris_dauksza_ai

Example Request

Use the Ollama API to send a generation request:

Request:

{
  "model": "chris_dauksza_ai",
  "prompt": "Write a Python function to calculate factorial using recursion.",
  "options": {
    "temperature": 0.2,
    "top_k": 20,
    "top_p": 0.6
  }
}

Response:

{
  "response": "def factorial(n):\n    if n == 0 or n == 1:\n        return 1\n    return n * factorial(n - 1)"
}

Advanced Configuration

The model supports a variety of parameters to customize its behavior:

Parameter Description Default Value
num_ctx Context window size 32768
temperature Creativity level 0.2
top_k Limits sampling to the top-k tokens 20
top_p Cumulative probability threshold 0.6
repeat_penalty Penalty for repeating tokens 1.2
mirostat Dynamic perplexity control 2

Development

This repository also includes API code that integrates with the model for generating, chatting, and embedding functionality. See the api directory for details.

Contributions

Contributions are welcome! Feel free to open an issue or submit a pull request if you have suggestions for improvement.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Contact

For questions or support, contact Chris Dauksza: - Email: chrisdauksza@gmail.com - LinkedIn: Chris Dauksza - Website: dauksza.fun


Empowering technology, one line of code at a time. “`