From d67be0c7de8fd717f648ff5aa2b1c169002c1697 Mon Sep 17 00:00:00 2001 From: myhloli Date: Fri, 14 Nov 2025 10:34:29 +0800 Subject: [PATCH] fix: add lmdeploy-engine parameters to compose.yaml for improved multi-GPU support --- docker/compose.yaml | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/docker/compose.yaml b/docker/compose.yaml index 8d37db59..b574a115 100644 --- a/docker/compose.yaml +++ b/docker/compose.yaml @@ -73,6 +73,9 @@ services: # parameters for vllm-engine # --data-parallel-size 2 # If using multiple GPUs, increase throughput using vllm's multi-GPU parallel mode # --gpu-memory-utilization 0.5 # If running on a single GPU and encountering VRAM shortage, reduce the KV cache size by this parameter, if VRAM issues persist, try lowering it further to `0.4` or below. + # parameters for lmdeploy-engine + # --dp 2 # If using multiple GPUs, increase throughput using lmdeploy's multi-GPU parallel mode + # --cache-max-entry-count 0.5 # If running on a single GPU and encountering VRAM shortage, reduce the KV cache size by this parameter, if VRAM issues persist, try lowering it further to `0.4` or below. ulimits: memlock: -1 stack: 67108864 @@ -98,12 +101,17 @@ services: command: --server-name 0.0.0.0 --server-port 7860 - --enable-vllm-engine true # Enable the vllm engine for Gradio # --enable-api false # If you want to disable the API, set this to false # --max-convert-pages 20 # If you want to limit the number of pages for conversion, set this to a specific number # parameters for vllm-engine + --enable-vllm-engine true # Enable the vllm engine for Gradio # --data-parallel-size 2 # If using multiple GPUs, increase throughput using vllm's multi-GPU parallel mode # --gpu-memory-utilization 0.5 # If running on a single GPU and encountering VRAM shortage, reduce the KV cache size by this parameter, if VRAM issues persist, try lowering it further to `0.4` or below. + # parameters for lmdeploy-engine + # !!!The lmdeploy and vLLM engines cannot be enabled simultaneously. Please ensure that at most only one engine is active at any given time.!!! + # --enable-lmdeploy-engine true # Enable the lmdeploy engine for Gradio + # --dp 2 # If using multiple GPUs, increase throughput using lmdeploy's multi-GPU parallel mode + # --cache-max-entry-count 0.5 # If running on a single GPU and encountering VRAM shortage, reduce the KV cache size by this parameter, if VRAM issues persist, try lowering it further to `0.4` or below. ulimits: memlock: -1 stack: 67108864