nanoAI-zoo

Concise, modular implementations of foundational AI models for research and experimentation across CV, LLM, VLM, and Generative AI.

Overview

nanoAI-zoo is a comprehensive collection of lightweight and modular implementations of foundational AI models. Designed for researchers and practitioners, it provides easy access to state-of-the-art architectures across various domains including Computer Vision (CV), Large Language Models (LLM), Vision-Language Models (VLM), and Generative AI.

Structure

The repository is organized into the following main directories:

• CV/: Implementations of foundational computer vision models.
• LLM/: Implementations of large language models.
• VLM/: Implementations of vision-language models.
• GenAI/: Implementations of generative AI models.
• utils/: Utility functions and tools to support model implementations.
• assets/: Pre-trained model weights and datasets.
• docs/: Documentation and tutorials.

Supported Models or Techniques

• Computer Vision (CV)

  • ResNet
  • Vision Transformer (ViT)
  • DINO
  • Swin Transformer
  • YOLOv7

• Large Language Models (LLM)

  • LoRA
  • Multi turn QA
  • inference with vllm

• Vision-Language Models (VLM)

  • Prevalent VLMs VQA-API(Qwen, Gemini)
  • CLIP
  • LLaVA
  • SigLIP
  • BLIP
  • Mobile-SAMv2
  • Grounding DINO

• Generative AI (GenAI)

  • GANs
  • VAEs
  • Diffusion
  • Flow-based Models

Acknowledgements

We would like to thank the open-source community for their contributions and inspiration in developing this repository. The repositories we reference are included in docs/references.md.

LICENSE

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