The modular core system in the VIRTUAL environment is designed to efficiently manage and integrate diverse contributions, enabling VIRTUALs to utilize various specialized cores (like cognitive, voice, visual) for different applications. This system is structured to maximize efficiency and adaptability in the value transfer between the cores and the VIRTUALs. Here's a breakdown of the contribution process and how these cores function:

Base Models:

  1. Community Contributions: Contributors from the community can provide generic models that serve as foundational Base Models. These models are versatile and can be applied across various contexts.

  2. Curated and Validated: The Base Models are curated and validated by the community, ensuring their quality and applicability.

  3. Foundation for Finetuning: These models are essential for further finetuning to meet specific requirements of a VIRTUAL. They serve as a starting point for more specialized model development.

Modular Cores:

  1. Specific to VIRTUAL Requirements: Modular cores are developed for specific functionalities within a VIRTUAL, like processing cognitive tasks, handling voice interactions, or managing visual inputs.

  2. Finetuning Base Models: To fit into a Modular Core, Base Models may require further finetuning. This process tailors the model to align with the specific needs and tasks of the VIRTUAL.

  3. Flexible and Interchangeable: The system allows VIRTUALs to easily swap between different models within these cores, facilitating adaptability and up-to-date performance.

  4. Governance and Future Development: The evolution or application of future cores is subject to decisions made within the VIRTUAL-specific governance community. This ensures that the development and integration of new cores are in line with the collective vision and standards of the community.

Contribute to Base Models

Like on Hugging Face, community members can enhance pre-existing models or submit new base models. This could involve improving the model's performance, expanding its dataset, or developing a completely new model with unique capabilities.

Once a model is submitted or enhanced, it goes through a Decentralized Autonomous Organization (DAO) proposal process. This involves the community reviewing and voting on the acceptance of the model. The DAO proposal ensures that the model aligns with the community's standards and requirements, maintaining a high level of quality and relevance.

Contribute to Modular Cores

Contributing to the Cognitive Core involves submitting text data to a Large Language Model (LLM). This data helps define and refine the central intelligence and personality of the VIRTUAL. The contributions could include domain-specific information, dialogues for personality development, or general knowledge to enhance the AI's understanding and response capabilities. The key is to provide rich, varied, and high-quality text data that can be used to train the LLM, making the VIRTUAL more knowledgeable, engaging, and effective in its interactions.

pageCognitive Core

Contribution towards Voice & Sound Core involves contributing any Voice Data or enhancements of the Voice Models.

pageVoice Core

Contribution towards Visual Core involves creation or improvements on Animation, 3D Visuals and Facial Expression.

pageVisual Core

Last updated