MiniMax-M3 GGUF model cannot be loaded - unknown model architecture 'minimax-m3'

[nfouka]

MiniMax-M3 GGUF model cannot be loaded - unknown model architecture 'minimax-m3'

Description

I am trying to load the MiniMax-M3 GGUF model with LocalAI v4.6.2, but the model fails to load because the embedded llama.cpp backend does not recognize the minimax-m3 architecture.

Environment

• LocalAI version: v4.6.2
• Backend: llama.cpp (internal loader)
• Model:
  MiniMaxAI.MiniMax-M3.IQ2_XS-00001-of-00008.gguf
• Model source:
  https://huggingface.co/DevQuasar/MiniMaxAI.MiniMax-M3-GGUF or https://huggingface.co/unsloth/MiniMax-M3-GGUF

Error message