Qwen3.8 Ik_llama

Prompt

this is on ubergarms qwen quant: git clone [https://github.com/ikawrakow/ik_llama.cpp.git](https://github.com/ikawrakow/ik_llama.cpp.git) cd ik_llama.cpp cmake -B build -DCMAKE_BUILD_TYPE=Release -DGGML_CUDA=ON -DGGML_CUDA_F16=ON cmake --build build --config Release -j $(nproc) # wget [https://huggingface.co/ubergarm/Qwen3.8-27B-GGUF/resolve/main/Qwen3.8-27B-MTP-IQ4_KS.gguf](https://huggingface.co/ubergarm/Qwen3.8-27B-GGUF/resolve/main/Qwen3.8-27B-MTP-IQ4_KS.gguf) model=/mnt/ai/models/ubergarm/Qwen3.8-27B-GGUF/Qwen3.8-27B-MTP-IQ4_KS.gguf # wget wget [https://huggingface.co/ggml-org/Qwen3.8-27B-GGUF/resolve/main/mmproj-Qwen3.8-27B-Q8_0.gguf](https://huggingface.co/ggml-org/Qwen3.8-27B-GGUF/resolve/main/mmproj-Qwen3.8-27B-Q8_0.gguf) mmproj=/mnt/ai/models/ubergarm/Qwen3.8-27B-GGUF/mmproj-Qwen3.8-27B-Q8_0.gguf CUDA_VISIBLE_DEVICES="0" \ ./build/bin/llama-server \ --model "$model" \ --alias "Qwen3.8-27B" \ -c 131072 \ -ctk q8_0 -ctv q8_0 \ -ctkd q8_0 -ctvd q8_0 \ --merge-qkv \ -muge \ -ngl 99 \ -t 1 \ -tb 1 \ -tm 16 \ --host 127.0.0.1 \ --port 8080 \ --parallel 1 \ --jinja \ --ctx-checkpoints 32 \ -cram 32768 \ --spec-type mtp:n_max=4,p_min=0.0 \ --no-mmproj-offload \ --mmproj "$mmproj" \ can you see if anything is missing here, check ik_llama github discussions commits etc if there is anything else missing and command line params in general that you would suggest me 3090rtx 5800x win11