Community ports MiniMax M3 EAGLE3 draft model to GGUF for llama.cpp

View on Reddit

The MiniMax M3 EAGLE3 draft model has been converted to GGUF format and is now compatible with llama.cpp. The conversion enables local inference on systems with 2x3090 GPUs and 128GB DDR4 RAM.