From 9c6fbd76be4edeb37298952a983427d6e3abdef1 Mon Sep 17 00:00:00 2001 From: Yueqian Lin <70319226+linyueqian@users.noreply.github.com> Date: Thu, 12 Mar 2026 12:30:20 -0400 Subject: [PATCH] Enable async_scheduling by default for Qwen3-TTS (#1853) Signed-off-by: linyueqian --- vllm_omni/model_executor/stage_configs/qwen3_tts.yaml | 4 ++-- vllm_omni/model_executor/stage_configs/qwen3_tts_batch.yaml | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/vllm_omni/model_executor/stage_configs/qwen3_tts.yaml b/vllm_omni/model_executor/stage_configs/qwen3_tts.yaml index 095aafe08..570ed710c 100644 --- a/vllm_omni/model_executor/stage_configs/qwen3_tts.yaml +++ b/vllm_omni/model_executor/stage_configs/qwen3_tts.yaml @@ -15,7 +15,7 @@ stage_args: scheduler_cls: vllm_omni.core.sched.omni_ar_scheduler.OmniARScheduler enforce_eager: false trust_remote_code: true - async_scheduling: false + async_scheduling: true enable_prefix_caching: false engine_output_type: latent gpu_memory_utilization: 0.3 @@ -50,7 +50,7 @@ stage_args: scheduler_cls: vllm_omni.core.sched.omni_generation_scheduler.OmniGenerationScheduler enforce_eager: true trust_remote_code: true - async_scheduling: false + async_scheduling: true enable_prefix_caching: false engine_output_type: audio gpu_memory_utilization: 0.2 diff --git a/vllm_omni/model_executor/stage_configs/qwen3_tts_batch.yaml b/vllm_omni/model_executor/stage_configs/qwen3_tts_batch.yaml index d73739109..5236668dc 100644 --- a/vllm_omni/model_executor/stage_configs/qwen3_tts_batch.yaml +++ b/vllm_omni/model_executor/stage_configs/qwen3_tts_batch.yaml @@ -19,7 +19,7 @@ stage_args: scheduler_cls: vllm_omni.core.sched.omni_ar_scheduler.OmniARScheduler enforce_eager: false trust_remote_code: true - async_scheduling: false + async_scheduling: true enable_prefix_caching: false engine_output_type: latent gpu_memory_utilization: 0.3 @@ -54,7 +54,7 @@ stage_args: scheduler_cls: vllm_omni.core.sched.omni_generation_scheduler.OmniGenerationScheduler enforce_eager: true trust_remote_code: true - async_scheduling: false + async_scheduling: true enable_prefix_caching: false engine_output_type: audio gpu_memory_utilization: 0.2