ModelReady is a comprehensive suite of tools designed to streamline data science processes for data scientists and AI practitioners. With a focus on quick setup, convenient launching, service management, model conversion, user prompts, and training automation, ModelReady makes it easy to adopt and manage data science tools in a wide range of scenarios.
- 🛠️ Installers: Simplify the setup of your data science environments with easy-to-use installation tools.
- 🚀 Launchers: Quickly get your data science applications up and running with our pre-configured launchers.
- 🔄 Systemd Services: Manage your data science system services efficiently and reduce overhead.
- 💡 Converters: Seamlessly convert models between different formats and frameworks, increasing versatility and adaptability.
- ❓ Prompts: User-friendly prompts guide you through various operations, making the toolkit more accessible and easy to use.
- 🏋️♀️ Training: Automate the training process to focus on fine-tuning your models, saving time and resources.
This project aims to provide a comprehensive suite of tools for managing data science processes, from model conversion and training to deploying AI applications. The toolkit is designed to work seamlessly with various machine learning frameworks and libraries.
Before installing the Data Science Toolkit, ensure that you have the following installed:
- Python 3.x (pip, venv)
- python3-is-python (if you need to)
- CUDA 10.2+ (recommend 11+, but >12 seems broken for some libraries)
- nvcc (usually comes with CUDA and is located in the CUDA_HOME/bin directory)
To get started with the Data Science Toolkit, clone this repository and install the required dependencies:
git clone git clone https://github.com/suparious/ModelReady.git
cd ModelReadyRefer to each feature's specific README for detailed instructions on how to use the provided scripts.
Installers provide a set of scripts that automatically create virtual Python environments and set up all necessary dependencies for machine learning applications. These scripts simplify the installation process and ensure a consistent environment across projects.
The launchers are a set of scripts designed to initiate various data science applications. These launchers streamline the deployment process and provide an efficient way to manage AI-related tasks.
This part of the toolkit offers a collection of systemd services tailored for working on remote machines. These services facilitate the deployment and management of data science tools on remote servers, enabling seamless integration into existing infrastructure.
This section contains a collection of scripts for converting between different model formats. These converters make it easy to work with various model formats and ensure compatibility with different machine learning libraries.
A collection of prompts is available to simplify the process of interacting with machine learning models. These prompts can be used as templates to generate custom text or to fine-tune models for specific use cases.
The training section includes a collection of scripts for training models. These scripts are designed to streamline the training process and make it easier to create and fine-tune custom machine learning models.
The Data Science Toolkit is designed to help manage data science processes in various scenarios. Refer to each feature's specific README for detailed instructions on how to use the provided scripts.
Contributions to the Data Science Toolkit are welcome. Please follow the contributing guidelines to ensure a smooth collaboration process.
This project is licensed under the GNU License. Please see the license file for more information.