pyproject.toml 4.66 KB
Newer Older
1
[build-system]
2
# Should be mirrored in requirements/build.txt
3
requires = [
4
    "cmake>=3.26.1",
5
    "ninja",
6
    "packaging>=24.2",
7
    "setuptools>=77.0.3,<81.0.0",
8
    "setuptools-scm>=8.0",
9
    "torch == 2.10.0",
10
    "wheel",
11
    "jinja2",
12
    "grpcio-tools==1.78.0",
13
14
]
build-backend = "setuptools.build_meta"
15

16
17
18
[project]
name = "vllm"
authors = [{name = "vLLM Team"}]
19
20
license = "Apache-2.0"
license-files = ["LICENSE"]
21
22
23
24
25
26
readme = "README.md"
description = "A high-throughput and memory-efficient inference and serving engine for LLMs"
classifiers = [
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
    "Programming Language :: Python :: 3.12",
27
    "Programming Language :: Python :: 3.13",
28
29
30
31
32
33
    "Intended Audience :: Developers",
    "Intended Audience :: Information Technology",
    "Intended Audience :: Science/Research",
    "Topic :: Scientific/Engineering :: Artificial Intelligence",
    "Topic :: Scientific/Engineering :: Information Analysis",
]
34
requires-python = ">=3.10,<3.14"
35
36
37
38
dynamic = [ "version", "dependencies", "optional-dependencies"]

[project.urls]
Homepage="https://github.com/vllm-project/vllm"
39
40
Documentation="https://docs.vllm.ai/en/latest/"
Slack="https://slack.vllm.ai/"
41
42
43
44

[project.scripts]
vllm = "vllm.entrypoints.cli.main:main"

45
46
[project.entry-points."vllm.general_plugins"]
lora_filesystem_resolver = "vllm.plugins.lora_resolvers.filesystem_resolver:register_filesystem_resolver"
47
lora_hf_hub_resolver = "vllm.plugins.lora_resolvers.hf_hub_resolver:register_hf_hub_resolver"
48

49
[tool.setuptools_scm]
50
# no extra settings needed, presence enables setuptools-scm
51
52
53

[tool.setuptools.packages.find]
where = ["."]
54
include = ["vllm*"]
55

56
[tool.ruff.lint.per-file-ignores]
57
"vllm/third_party/**" = ["ALL"]
58
59
"vllm/version.py" = ["F401"]
"vllm/_version.py" = ["ALL"]
60
61
62
63
# Exclude generated protobuf files
"vllm/grpc/*_pb2.py" = ["ALL"]
"vllm/grpc/*_pb2_grpc.py" = ["ALL"]
"vllm/grpc/*_pb2.pyi" = ["ALL"]
64

65
66
67
68
69
70
71
[tool.ruff.lint]
select = [
    # pycodestyle
    "E",
    # Pyflakes
    "F",
    # pyupgrade
72
    "UP",
73
74
    # flake8-bugbear
    "B",
75
76
    # flake8-implicit-str-concat
    "ISC",
77
78
79
    # flake8-simplify
    "SIM",
    # isort
80
    "I",
81
    # flake8-logging-format
82
    "G",
83
84
85
86
87
88
]
ignore = [
    # star imports
    "F405", "F403",
    # lambda expression assignment
    "E731",
89
90
    # zip without `strict=`
    "B905",
91
92
    # Loop control variable not used within loop body
    "B007",
93
94
    # f-string format
    "UP032",
95
]
96

97
98
99
[tool.ruff.format]
docstring-code-format = true

100
[tool.mypy]
101
plugins = ['pydantic.mypy']
102
ignore_missing_imports = true
103
check_untyped_defs = true
104
follow_imports = "silent"
105

106
107
[tool.pytest.ini_options]
markers = [
108
    "slow_test",
109
    "skip_global_cleanup",
110
    "core_model: enable this model test in each PR instead of only nightly",
111
    "hybrid_model: models that contain mamba layers (including pure SSM and hybrid architectures)",
112
    "cpu_model: enable this model test in CPU tests",
113
    "cpu_test: mark test as CPU-only test",
114
115
    "split: run this test as part of a split",
    "distributed: run this test only in distributed GPU tests",
116
    "optional: optional tests that are automatically skipped, include --optional to run them",
117
]
118

119
[tool.ty.src]
Aaron Pham's avatar
Aaron Pham committed
120
121
122
123
respect-ignore-files = true

[tool.ty.environment]
python = "./.venv"
124
125
126

[tool.typos.files]
# these files may be written in non english words
Jiayi Yan's avatar
Jiayi Yan committed
127
128
129
130
131
extend-exclude = ["tests/models/fixtures/*", "tests/prompts/*", "tests/tokenizers_/*",
    "benchmarks/sonnet.txt", "tests/lora/data/*", "examples/pooling/token_embed/*", "build/*",
    "vllm/third_party/*", "vllm/entrypoints/serve/instrumentator/static/*", "tests/entrypoints/openai/test_transcription_validation.py",
    "docs/governance/process.md", "tests/v1/engine/test_fast_incdec_prefix_err.py", ".git/*"]
ignore-hidden = false
132
133

[tool.typos.default]
Jiayi Yan's avatar
Jiayi Yan committed
134
extend-ignore-identifiers-re = [".*[Uu][Ee][0-9][Mm][0-9].*"]
135
136
137
138
139
140

[tool.typos.default.extend-identifiers]
bbc5b7ede = "bbc5b7ede"
NOOPs = "NOOPs"
nin_shortcut = "nin_shortcut"
cudaDevAttrMaxSharedMemoryPerBlockOptin = "cudaDevAttrMaxSharedMemoryPerBlockOptin"
Jiayi Yan's avatar
Jiayi Yan committed
141

142
depthwise_seperable_out_channel = "depthwise_seperable_out_channel"
Jiayi Yan's avatar
Jiayi Yan committed
143
144
145
146
147
148
149
150
pard_token = "pard_token"
ptd_token_id = "ptd_token_id"
ser_de = "ser_de"
shared_memory_per_block_optin = "shared_memory_per_block_optin"
FoPE = "FoPE"
k_ot = "k_ot"
view_seperator = "view_seperator"
inverse_std_variences = "inverse_std_variences"
151
152
153

[tool.typos.default.extend-words]
iy = "iy"
154
indx = "indx"
155
156
157
158
159
# intel cpu features
tme = "tme"
dout = "dout"
Pn = "Pn"
arange = "arange"
Jiayi Yan's avatar
Jiayi Yan committed
160
161
162
163
164
165
166
thw = "thw"
subtile = "subtile"
HSA = "HSA"
setp = "setp"
CPY = "CPY"
thr = "thr"
Thr = "Thr"
167
168
PARD = "PARD"
pard = "pard"
169
AKS = "AKS"
170
ba = "ba"
Jiayi Yan's avatar
Jiayi Yan committed
171
fo = "fo"
172
nd = "nd"
Jiayi Yan's avatar
Jiayi Yan committed
173
174
eles = "eles"
datas = "datas"
175
176

[tool.uv]
177
no-build-isolation-package = ["torch"]