From b2de3b17eee6391144fa20ae851278141163de20 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Sebastian=20Kr=C3=BCger?= Date: Sun, 30 Nov 2025 20:16:00 +0100 Subject: [PATCH] fix: adjust VRAM allocation for concurrent Llama+BGE MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Llama: 85% GPU, 8K context (model needs ~15GB base) - BGE: 10% GPU (1.3GB model) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude --- services/vllm/config_bge.yaml | 2 +- services/vllm/config_llama.yaml | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/services/vllm/config_bge.yaml b/services/vllm/config_bge.yaml index aba81c0..1384ad5 100644 --- a/services/vllm/config_bge.yaml +++ b/services/vllm/config_bge.yaml @@ -2,6 +2,6 @@ model: BAAI/bge-large-en-v1.5 host: "0.0.0.0" port: 8002 uvicorn-log-level: "info" -gpu-memory-utilization: 0.15 +gpu-memory-utilization: 0.10 dtype: float16 task: embed diff --git a/services/vllm/config_llama.yaml b/services/vllm/config_llama.yaml index e0331fa..b08af74 100644 --- a/services/vllm/config_llama.yaml +++ b/services/vllm/config_llama.yaml @@ -2,8 +2,8 @@ model: meta-llama/Llama-3.1-8B-Instruct host: "0.0.0.0" port: 8001 uvicorn-log-level: "info" -gpu-memory-utilization: 0.70 -max-model-len: 16384 +gpu-memory-utilization: 0.85 +max-model-len: 8192 dtype: auto enforce-eager: false enable-auto-tool-choice: true