Overview

Cydonia-v4.1-MS3.2-Magnum-Diamond-24B is a 24 billion parameter language model developed by knifeayumu, representing an updated iteration in the Cydonia series. This model was constructed using the SLERP merge method within mergekit, combining two distinct base models: TheDrummer/Cydonia-24B-v4.1 and Doctor-Shotgun/MS3.2-24B-Magnum-Diamond.

Merge Details

The model's architecture is a direct result of merging specific pre-trained language models. The configuration used for this merge involved:

• Base Model: TheDrummer/Cydonia-24B-v4.1
• Merged Models:
  • TheDrummer/Cydonia-24B-v4.1
  • Doctor-Shotgun/MS3.2-24B-Magnum-Diamond
• Merge Method: SLERP
• Data Type: bfloat16

This merge strategy aims to integrate the strengths of its constituent models, offering an updated version for users interested in the Cydonia model line. The model supports a context length of 32768 tokens.

Additional Information

The model card also references the use of image-to-video generation tools like Wan-AI/Wan2.2-I2V-A14B with lightx2v/Wan2.2-Lightning and pollockjj/ComfyUI-MultiGPU nodes, indicating potential applications or demonstrations involving multimedia, though the model itself is a language model.