Tag
Vision support for the Minimax-M3 model has been merged into the llama.cpp project, enabling multimodal inference for this model locally.