This repository curates foundational and multimodal AI models, serving as a comprehensive resource for researchers and developers exploring advanced AI capabilities. It provides a structured overview of cutting-edge models that integrate vision, language, and audio, enabling a wide array of downstream tasks.

How It Works

The list categorizes models by their supported modalities (vision, language, audio) and primary tasks, such as object detection, segmentation, and text-to-audio generation. Each entry includes key details like publication date, associated papers, code repositories, and example usage, facilitating quick evaluation and adoption.

Quick Start & Requirements

This is a curated list, not a runnable codebase. Users will need to refer to individual model repositories for installation and execution instructions.

Highlighted Details

• Covers models from late 2021 to early 2024, showcasing recent advancements.
• Includes models with diverse modality combinations: vision-only, vision-language, and audio-language.
• Features prominent models like YOLO-World, Depth Anything, Segment Anything, LLaVA, and Whisper.
• Lists tasks ranging from zero-shot object detection and segmentation to image captioning and speech recognition.

Maintenance & Community

The repository is community-driven, actively seeking contributions for new models and improvements via issues and pull requests.

Licensing & Compatibility

Licensing information is not provided for the curated list itself. Users must consult the individual model repositories for their respective licenses and compatibility.

Limitations & Caveats

This is a reference list and does not provide a unified API or execution environment. Users must independently manage dependencies and setup for each model.