From ed37d8e7a27a51051804abc87dacf8c10fb10201 Mon Sep 17 00:00:00 2001 From: akshatvishu <33392262+akshatvishu@users.noreply.github.com> Date: Wed, 17 Jun 2026 18:19:56 +0530 Subject: [PATCH] [Docs] Fix hallucinated stage CLI flags in documentation and comments (#4512) Signed-off-by: akshatvishu --- docs/user_guide/examples/online_serving/glm_image.md | 2 +- vllm_omni/deploy/mimo_audio.yaml | 4 ++-- vllm_omni/model_executor/stage_configs/mimo_audio.yaml | 4 ++-- 3 files changed, 5 insertions(+), 5 deletions(-) diff --git a/docs/user_guide/examples/online_serving/glm_image.md b/docs/user_guide/examples/online_serving/glm_image.md index 410b1180a..ea404faa9 100644 --- a/docs/user_guide/examples/online_serving/glm_image.md +++ b/docs/user_guide/examples/online_serving/glm_image.md @@ -19,7 +19,7 @@ both stages on a single GPU, override per stage: ```bash vllm serve zai-org/GLM-Image --omni --port 8091 \ - --stage-0-devices 0 --stage-1-devices 0 + --stage-overrides '{"0": {"devices": "0"}, "1": {"devices": "0"}}' ``` ## API Calls diff --git a/vllm_omni/deploy/mimo_audio.yaml b/vllm_omni/deploy/mimo_audio.yaml index 6ebe59209..a50632964 100644 --- a/vllm_omni/deploy/mimo_audio.yaml +++ b/vllm_omni/deploy/mimo_audio.yaml @@ -4,8 +4,8 @@ # Default mode is async-chunk streaming on a single GPU (both stages on # device 0) through SharedMemoryConnector. For the legacy 2-GPU sync # pipeline (stage 1 on a second card), pass ``--no-async-chunk -# --stage-1-devices 1 --stage-1-max-model-len 18192 -# --stage-1-max-num-batched-tokens 18192`` to the serve CLI. +# --stage-overrides '{"1": {"devices": "1", "max_model_len": 18192, "max_num_batched_tokens": 18192}}'`` +# to the serve CLI. async_chunk: true dtype: bfloat16 diff --git a/vllm_omni/model_executor/stage_configs/mimo_audio.yaml b/vllm_omni/model_executor/stage_configs/mimo_audio.yaml index 493c05f00..b81f27fe4 100644 --- a/vllm_omni/model_executor/stage_configs/mimo_audio.yaml +++ b/vllm_omni/model_executor/stage_configs/mimo_audio.yaml @@ -4,8 +4,8 @@ # Default mode is async-chunk streaming on a single GPU (both stages on # device 0) through SharedMemoryConnector. For the legacy 2-GPU sync # pipeline (stage 1 on a second card), pass ``--no-async-chunk -# --stage-1-devices 1 --stage-1-max-model-len 18192 -# --stage-1-max-num-batched-tokens 18192`` to the serve CLI. +# --stage-overrides '{"1": {"devices": "1", "max_model_len": 18192, "max_num_batched_tokens": 18192}}'`` +# to the serve CLI. async_chunk: true dtype: bfloat16