Update to Lemonade v10.5.1, then: ``` Get the model lemonade pull Qwen3.6-27B-MTP-GGUF Get ROCm 7.13 lemonade backends install llamacpp:rocm Load the model (MTP args auto-applied) lemonade load Qwen3.6-27B-MTP-GGUF --llamacpp rocm --ctx-size 0 ``` Shown in the video taking a look in the mirror with
