From 72cb4f661b6ba138c584856202cea6ef170ff02c Mon Sep 17 00:00:00 2001 From: tzhouam Date: Thu, 12 Mar 2026 13:57:52 +0000 Subject: [PATCH] Adjust GPU memory utilization and device settings in qwen2_5_omni_ci.yaml for improved performance Signed-off-by: tzhouam --- tests/e2e/stage_configs/qwen2_5_omni_ci.yaml | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tests/e2e/stage_configs/qwen2_5_omni_ci.yaml b/tests/e2e/stage_configs/qwen2_5_omni_ci.yaml index c11de3e5d..188f219ec 100644 --- a/tests/e2e/stage_configs/qwen2_5_omni_ci.yaml +++ b/tests/e2e/stage_configs/qwen2_5_omni_ci.yaml @@ -47,7 +47,7 @@ stage_args: max_model_len: 16384 max_num_batched_tokens: 16384 max_num_seqs: 1 - gpu_memory_utilization: 0.9 + gpu_memory_utilization: 0.4 skip_mm_profiling: true enforce_eager: true trust_remote_code: true @@ -67,14 +67,14 @@ stage_args: - stage_id: 2 runtime: process: true - devices: "2" # Example: use a different GPU than the previous stage; use "0" if single GPU + devices: "1" # Example: use a different GPU than the previous stage; use "0" if single GPU max_batch_size: 1 engine_args: model_stage: code2wav model_arch: Qwen2_5OmniForConditionalGeneration worker_type: generation scheduler_cls: vllm_omni.core.sched.omni_generation_scheduler.OmniGenerationScheduler - gpu_memory_utilization: 0.9 #increase the gpu memory utilization to enable the test on H800 + gpu_memory_utilization: 0.5 #increase the gpu memory utilization to enable the test on H800 enforce_eager: true trust_remote_code: true enable_prefix_caching: false