Skip to main content
The ModelMergeCosmos14B node merges two AI models using a block-based approach specifically designed for the Cosmos 14B model architecture. It allows you to blend different components of the models by adjusting weight values between 0.0 and 1.0 for each model block and embedding layer.

Inputs

All weight parameters accept values from 0.0 to 1.0 in steps of 0.01, and every one of them defaults to 1.0. The block parameters cover 36 transformer blocks, numbered blocks.block0. through blocks.block35..

Outputs

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): 1d1e5dc176643f577723bb0bb9375748a392a6fafa5c9e5e78ef4c4d8289f77c