Skip to content

kyegomez/CELESTIAL-1

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

35 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Multi-Modality

CELESTIAL-1: A Model for Processing, Understanding, and Reasoning in Unlimited Modalities

GitHub issues GitHub forks GitHub stars GitHub license Share on Twitter Share on Facebook Share on LinkedIn Discord Share on Reddit Share on Hacker News Share on Pinterest Share on WhatsApp

CELESTIAL-1 is an open source model designed to provide a general representation for processing, understanding, and reasoning across an unlimited range of modalities.

CELESTIAL-1 can operate in vision, audio, audio-language, and vision-language tasks without any prior initialization.

This model will change everything for our species, we'll able to automate any digital task. And, we need your help. Contribute to the project board here:

Install

pip install cstx

Usage

Features

CELESTIAL-1 boasts a wide range of features designed to enhance its versatility and usefulness. Some of these features include:

  • Unlimited Modalities: The model can process and understand an unlimited number of modalities, increasing its applicability in a wide range of tasks.
  • Reasoning in Context: CELESTIAL-1 is capable of reasoning in context across multiple modalities, thus improving its understanding and representation of complex scenarios.
  • Generates Any Modality: With the ability to generate any modality, CELESTIAL-1 can cater to a wider variety of applications.
  • Scaling-Friendly Architecture: The architecture of CELESTIAL-1 is designed to be friendly to scaling, which allows it to efficiently handle larger and more complex tasks.
  • Modality-Agnostic Tasks: The model can perform tasks that are agnostic to the modality, further widening its range of uses.
  • Strong Emergent Zero-Shot Retrieval Capability: CELESTIAL-1 has an impressive ability to retrieve information in a zero-shot manner, an essential feature for any advanced AI model.

Contribution

Contributing to CELESTIAL-1 is a rewarding experience, whether you're a novice developer or an AI expert. Here's a simple guide to get you started:

  1. Understand the Project: Familiarize yourself with CELESTIAL-1's mission and features. Check out the project board: Contribute to the project board.

  2. Choose a Task: Select a task that resonates with your skill set and passion. It could range from documentation, coding, testing, or even discussions.

  3. Fork the Repository: Create a personal fork of the CELESTIAL-1 repository on GitHub.

  4. Clone Your Fork:

    git clone https://github.com/<your-username>/Celestial-1.git
  5. Branching: Always create a new branch for each task. This keeps things organized.

    git checkout -b feature/your-branch-name
  6. Code, Document, Discuss: Dive in! Make your changes, write/document, or participate in discussions.

  7. Commit and Push: After making your changes, commit them and push them to your fork.

    git commit -m "Your concise commit message"
    git push origin feature/your-branch-name
  8. Create a Pull Request: Go to the original CELESTIAL-1 repository on GitHub and create a pull request. Fill out the provided template with as much detail as possible.

  9. Wait for Review: Maintainers or other contributors will review your changes. They might ask for edits or further information. Collaboration is key!

  10. Celebrate: Once your contribution is merged, celebrate! You've just made an impactful contribution to an incredible project.

Tips:

  • Always sync your fork with the main repository before starting on a new feature.
  • Clearly document any changes or additions you make.
  • Participate actively in discussions and respect all contributors.

Remember, every contribution, no matter how small, is valuable. Your insights, skills, and time are greatly appreciated. Let's make CELESTIAL-1 remarkable together!


References

For more information about the ONE-PEACE architecture that CELESTIAL-1 is based on, you can refer to the following resources:

  • ONE-PEACE Paper: Link
  • ONE-PEACE Demo: Link
  • ONE-PEACE Checkpoints: Link

Vicuna,

ImageBind,

Stable Diffusion,

AudioLDM Zeroscope.

PandaGPT,

VPGTrans,

GILL,

CoDi,

Video-LLaMA,

and MiniGPT-4


License

CELESTIAL-1 is released under the Apache 2.0 license. For more information about what this license entails, please refer to the LICENSE file.