A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 12, 23:30 UTC
Speed up macOS smoke test (#28954)
Signed-off-by: Michael Goin <mgoin64@gmail.com> Signed-off-by: mgoin <mgoin64@gmail.com>
Michael Goin committed
Nov 19, 2025 at 01:46 UTC
a4511e38db375a85b4dd784c2c38528747288f46
1 file changed
+3
-4
.github/workflows/macos-smoke-test.yml
+3
-4
@@ -9,7 +9,7 @@ on:
9
jobs:
10
macos-m1-smoke-test:
11
runs-on: macos-latest
12
- timeout-minutes: 20
12
+ timeout-minutes: 30
13
14
steps:
15
- uses: actions/checkout@v4
@@ -37,15 +37,14 @@ jobs:
37
- name: Verify installation
38
run: |
39
python -c "import vllm; print(f'vLLM version: {vllm.__version__}')"
40
- python -c "import torch; print(f'PyTorch: {torch.__version__}')"
40
41
- name: Smoke test vllm serve
43
- timeout-minutes: 10
42
run: |
43
# Start server in background
44
vllm serve Qwen/Qwen3-0.6B \
47
- --max-model-len=2048 \
45
+ --max-model-len=2K \
46
--load-format=dummy \
47
+ --hf-overrides '{"num_hidden_layers": 2}' \
48
--enforce-eager \
49
--port 8000 &
50