AI_MK_Toolkit

AI_MK_Toolkit is a production-ready Python package that provides a unified, consistent interface for accessing over 205 machine learning, deep learning, natural language processing (NLP), computer vision (CV), and reinforcement learning (RL) models and tasks from leading open-source frameworks.

Rather than reimplementing algorithms, AI_MK_Toolkit acts as a lightweight abstraction layer that simplifies the use of multiple AI frameworks while preserving the functionality, performance, and flexibility of their official implementations.

Python Requirement: Python 3.10 or later

Supported Platforms: Windows, Linux, and macOS

---

🚀 Why AI_MK_Toolkit?

Modern AI development often requires switching between multiple frameworks, each with its own APIs, import patterns, dependency management, and usage conventions. AI_MK_Toolkit removes this complexity by providing a single, intuitive, and consistent interface across popular AI ecosystems, allowing developers to focus on building solutions instead of framework-specific boilerplate.

---

✨ Key Features

• Unified API – Access models and tasks from multiple AI frameworks through a consistent interface.
• 205+ Supported Entries – Includes 51 Machine Learning models, 67 Deep Learning models, 54 NLP models, 19 Reinforcement Learning algorithms, and 14 Computer Vision task APIs.
• Lazy Loading – Optional dependencies are imported only when required, reducing startup time and installation overhead.
• Framework Agnostic – Use models from libraries such as scikit-learn, PyTorch, TensorFlow, XGBoost, LightGBM, Hugging Face Transformers, Stable-Baselines3, and more through a common API.
• Production Ready – Built with type hints, comprehensive error handling, and PEP 561 (py.typed) support.
• Modular Installation – Install only the dependencies needed for your specific use case using optional extras.
• Transparent Attribution – Clearly documents the original projects and their licenses, ensuring proper acknowledgment of all underlying open-source libraries.
• Developer Friendly – Simplifies experimentation, prototyping, education, and production workflows by minimizing framework-specific code.

---

📖 About This Project

AI-MK-Toolkit is a production-ready Python package that provides a unified, consistent interface to over 205 machine learning, deep learning, natural language processing, computer vision, reinforcement learning, and large language model (LLM) implementations from widely adopted open-source frameworks.

Rather than reimplementing algorithms, AI-MK-Toolkit serves as a lightweight abstraction layer that enables developers to access models from multiple ecosystems through a single, intuitive API while preserving the functionality and performance of the underlying libraries.

---

Key Principles

• Unified API – Access models from multiple frameworks using a consistent interface.
• No Algorithm Reimplementation – Uses the official implementations provided by libraries such as scikit-learn, PyTorch, TensorFlow, XGBoost, LightGBM, Hugging Face Transformers, Stable-Baselines3, and others.
• Modular Installation – Install only the dependencies required for your specific use case through optional extras.
• Transparent Attribution – Properly acknowledges and respects the licenses of all underlying open-source projects.
• Developer-Friendly – Simplifies experimentation, prototyping, and production development by reducing framework-specific boilerplate.

Important: AI-MK-Toolkit is an interface and integration library. It does not claim ownership of, modify, or replace the original machine learning algorithms or pretrained models provided by the respective open-source projects. All credit for the underlying implementations belongs to their original authors and maintainers.

---

📦 Installation

Base Installation

pip install ai-mk-toolkit

Deep Learning

pip install ai-mk-toolkit[dl]

Natural Language Processing

pip install ai-mk-toolkit[nlp]

Reinforcement Learning

pip install ai-mk-toolkit[rl]

Complete Installation

pip install ai-mk-toolkit[full]

---

🚀 Quick Start

Machine Learning

from ai_mk_toolkit import ml

model = ml.RandomForest()
model.fit(X_train, y_train)

predictions = model.predict(X_test)

Deep Learning

from ai_mk_toolkit import dl

model = dl.ResNet50(pretrained=True)

Natural Language Processing

from ai_mk_toolkit import nlp

bert = nlp.BERT()
tokenizer = nlp.Tokenizer()

Reinforcement Learning

from ai_mk_toolkit import rl

agent = rl.DQN(policy="MlpPolicy", env=env)

Computer Vision

from ai_mk_toolkit import cv

detector = cv.ObjectDetection()
classifier = cv.ImageClassification()

---

📊 Package Overview

Module	Description
ML	51 classical machine learning models
DL	67 deep learning model wrappers
NLP	54 NLP models and task APIs
RL	19 reinforcement learning algorithms
CV	14 computer vision task APIs

Total: 205+ unified model wrappers and task interfaces

---

🏗 Package Structure

ai_mk_toolkit/
├── ml/
├── dl/
├── nlp/
├── rl/
├── cv/
├── automl/
├── datasets/
├── preprocessing/
├── metrics/
├── explainability/
├── visualization/
├── deployment/
├── plugins/
├── backends/
└── core/

---

✅ Quality

• Production-ready package structure
• Type annotated (PEP 561)
• Comprehensive automated test suite
• Wheel and source distributions
• Lazy-loading architecture
• Clean wrapper design (no algorithm reimplementation)

---

🧪 Testing

pytest

pytest -v

---

📄 License

AI_MK_Toolkit is distributed under the MIT License.

The package wraps existing open-source libraries, each of which retains its own license.

Important: Some optional dependencies have copyleft licenses (for example, Ultralytics (AGPL-3.0) and Gensim (LGPL-2.1+)). Users are responsible for complying with the licenses of any optional dependencies they choose to install.

---

🙏 Acknowledgments

AI_MK_Toolkit builds upon the excellent work of the open-source community.

Major frameworks include:

• scikit-learn
• NumPy
• PyTorch
• torchvision
• TensorFlow / Keras
• Hugging Face Transformers
• XGBoost
• LightGBM
• CatBoost
• Stable-Baselines3
• timm
• spaCy
• NLTK
• Gensim
• sentence-transformers
• Ultralytics

Important: All trademarks, models, algorithms, and implementations remain the intellectual property of their respective authors and maintainers.

---

🤝 Contributing

Contributions, feature requests, and bug reports are welcome.

Before contributing:

• Follow the existing project structure.
• Document any new dependency and its license.
• Prefer permissive licenses (MIT, BSD, Apache-2.0).
• Keep optional dependencies modular whenever possible.

---

⚠ Disclaimer

AI_MK_Toolkit is a wrapper library, not a machine learning framework.

Important: It provides a unified Python interface to official implementations from existing open-source projects and does not claim ownership or authorship of the underlying algorithms or pretrained models. All credit belongs to the original authors and maintainers of those projects.