Inquiry: Feasibility of Merging Our SOTA Video Editing Model (DiT-MoE Based）

[SUISMOI]

Dear Community Members,
Greetings! Our team has developed a state-of-the-art (SOTA) model for video editing tasks. Before officially starting the development work for merging this model into the diffusers repository, we are reaching out to consult the community on relevant rules and opinions. Thank you for your valuable time.

Core Project Information
Project： mammothmoda 2.5
Project repository URL: https://mammothmoda2.github.io/
Model Positioning: It is the first SOTA video editing model built on the DiT-MoE architecture in the industry, focusing on optimizing performance and effects for video editing scenarios;
Technical Foundation: The backbone of our model is adapted from the "wan" project, but we have made substantial modifications to its core logic and architecture, with significant changes.

Core Questions to Clarify
Given the above background, does this model meet the merging criteria of the diffusers repository? We hope to confirm the community's acceptance before development to avoid ineffective work;
If the community approves the merge, we will inevitably need to reference parts of the "wan" code during development. Would it be permissible to directly import these parts of the "wan" project's code into our implementation?

We highly value the community's code specifications and open-source guidelines, and aim to ensure all operations comply with community requirements from the initial stage of development. We look forward to your feedback and suggestions.

Thank you!