From a88c427ba9651f55909711311317a4403fc0e859 Mon Sep 17 00:00:00 2001 From: dschlueter Date: Mon, 6 Jul 2026 18:47:50 +0200 Subject: [PATCH] config: model profiles inherit the standard port/container Replace the stale example profiles (qwen35->8002, deepseek->8003, pointing at non-existent paths) with profiles for the actually installed models. They now override ONLY model_path and inherit host_port=8001, container_name and alias from [default] -> one model at a time on the standard port; --start/--change swaps it. Distinct port/container is only needed for concurrent operation. Co-Authored-By: Claude Opus 4.8 --- llama.cpp.config | 28 +++++++++++++--------------- 1 file changed, 13 insertions(+), 15 deletions(-) diff --git a/llama.cpp.config b/llama.cpp.config index 6ad77aa..09cee2a 100644 --- a/llama.cpp.config +++ b/llama.cpp.config @@ -57,22 +57,20 @@ max_tokens = 32768 # chat_temperature leer lassen -> Server-Temp (temp oben) gilt. chat_temperature = -# Example second model profile, pinned to the second GPU (e.g. RTX 3090 #2) -# with its own port and container name so it can run alongside [default]. -[model.qwen35] -model_path = qwen3/Qwen3-35B-A3B-Q4_K_M.gguf -model_alias = qwen35 -container_name = llama_cpp_qwen35 -host_port = 8002 -gpu_device = 1 +# Modell-Profile: erben ALLES aus [default] (Port 8001, Container +# llama_cpp_server, Alias default_llm, gpu_device 1, ctx_size 262144) und +# überschreiben NUR das Modell. Es läuft also immer ein Modell auf dem +# Standard-Port; --start/--change tauscht es aus. Wer Modelle GLEICHZEITIG +# betreiben will, gibt dem Profil zusätzlich einen eigenen container_name + +# host_port (Muster siehe llama.cpp.config.example). +[model.carnice] +model_path = models/qwen3/Carnice-Qwen3.6-MoE-35B-A3B-Q4_K_M.gguf -[model.deepseek] -model_path = deepseek/deepseek-r1-q4.gguf -model_alias = deepseek -container_name = llama_cpp_deepseek -host_port = 8003 -gpu_device = 1 -ctx_size = 131072 +[model.qwen27] +model_path = models/qwen3/Qwen3.6-27B-Uncensored-HauhauCS-Aggressive-IQ4_XS.gguf + +[model.qwopus] +model_path = models/qwen3/Qwopus3.6-35B-A3B-v1-Q4_K_M.gguf [prompt.concise] system_prompt = Du antwortest kurz, präzise und technisch.