Skip to content

GitLab

Unverified Commit 31f31e8e authored by Ziqi Fan's avatar Ziqi Fan Committed by GitHub
Browse files

fix: remove small --gpu-memory-utilization to avoid OOM due to vllm upgrade (#4899) 


Signed-off-by: default avatarZiqi Fan <ziqif@nvidia.com>
parent 5e96c9a8
Show whitespace changes
Inline Side-by-side
Showing with 0 additions and 10 deletions
examples/backends/vllm/deploy/agg_kvbm.yaml
View file @ 31f31e8e
......@@ -40,8 +40,6 @@ spec:
args:
- --model
- Qwen/Qwen3-8B
- --gpu-memory-utilization
- "0.45"
- --max-model-len
- "32000"
- --enforce-eager
......
examples/backends/vllm/deploy/disagg_kvbm.yaml
View file @ 31f31e8e
......@@ -33,8 +33,6 @@ spec:
args:
- --model
- Qwen/Qwen3-8B
- --gpu-memory-utilization
- "0.3"
- --max-model-len
- "32000"
- --enforce-eager
......@@ -65,8 +63,6 @@ spec:
- --model
- Qwen/Qwen3-8B
- --is-prefill-worker
- --gpu-memory-utilization
- "0.3"
- --max-model-len
- "32000"
- --enforce-eager
......
examples/backends/vllm/deploy/disagg_kvbm_2p2d.yaml
View file @ 31f31e8e
......@@ -33,8 +33,6 @@ spec:
args:
- --model
- Qwen/Qwen3-8B
- --gpu-memory-utilization
- "0.3"
- --max-model-len
- "32000"
- --enforce-eager
......@@ -65,8 +63,6 @@ spec:
- --model
- Qwen/Qwen3-8B
- --is-prefill-worker
- --gpu-memory-utilization
- "0.3"
- --max-model-len
- "32000"
- --enforce-eager
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment