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 6, 18:15 UTC
[CI] Shard Multi-Modal Models (Standard) into 4 parallel jobs (#37014)
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Kevin H. Luu committed
Mar 14, 2026 at 01:26 UTC
600a039f572ac28128750f0463af428c5a260f1a
1 file changed
+48
-4
.buildkite/test_areas/models_multimodal.yaml
+48
-4
@@ -2,15 +2,59 @@ group: Models - Multimodal
2
depends_on:
3
- image-build
4
steps:
5
-- label: Multi-Modal Models (Standard) # 60min
6
- timeout_in_minutes: 80
5
+- label: "Multi-Modal Models (Standard) 1: qwen2"
6
+ timeout_in_minutes: 45
7
source_file_dependencies:
8
- vllm/
9
- tests/models/multimodal
10
commands:
11
- pip install git+https://github.com/TIGER-AI-Lab/Mantis.git
12
- - pip freeze | grep -E 'torch'
13
- - pytest -v -s models/multimodal -m core_model --ignore models/multimodal/generation/test_whisper.py --ignore models/multimodal/processing
12
+ - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "qwen2"
13
+ - pytest -v -s models/multimodal/generation/test_ultravox.py -m core_model
14
+ mirror:
15
+ amd:
16
+ device: mi325_1
17
+ depends_on:
18
+ - image-build-amd
19
+
20
+- label: "Multi-Modal Models (Standard) 2: qwen3 + gemma"
21
+ timeout_in_minutes: 45
22
+ source_file_dependencies:
23
+ - vllm/
24
+ - tests/models/multimodal
25
+ commands:
26
+ - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git
27
+ - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "qwen3 or gemma"
28
+ - pytest -v -s models/multimodal/generation/test_qwen2_5_vl.py -m core_model
29
+ mirror:
30
+ amd:
31
+ device: mi325_1
32
+ depends_on:
33
+ - image-build-amd
34
+
35
+- label: "Multi-Modal Models (Standard) 3: llava + qwen2_vl"
36
+ timeout_in_minutes: 45
37
+ source_file_dependencies:
38
+ - vllm/
39
+ - tests/models/multimodal
40
+ commands:
41
+ - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git
42
+ - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "not qwen2 and not qwen3 and not gemma"
43
+ - pytest -v -s models/multimodal/generation/test_qwen2_vl.py -m core_model
44
+ mirror:
45
+ amd:
46
+ device: mi325_1
47
+ depends_on:
48
+ - image-build-amd
49
+
50
+- label: "Multi-Modal Models (Standard) 4: other + whisper"
51
+ timeout_in_minutes: 45
52
+ source_file_dependencies:
53
+ - vllm/
54
+ - tests/models/multimodal
55
+ commands:
56
+ - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git
57
+ - pytest -v -s models/multimodal -m core_model --ignore models/multimodal/generation/test_common.py --ignore models/multimodal/generation/test_ultravox.py --ignore models/multimodal/generation/test_qwen2_5_vl.py --ignore models/multimodal/generation/test_qwen2_vl.py --ignore models/multimodal/generation/test_whisper.py --ignore models/multimodal/processing
58
- cd .. && VLLM_WORKER_MULTIPROC_METHOD=spawn pytest -v -s tests/models/multimodal/generation/test_whisper.py -m core_model # Otherwise, mp_method="spawn" doesn't work
59
mirror:
60
amd: