This repository has been archived by the owner on Oct 11, 2024. It is now read-only.
forked from vllm-project/vllm
-
Notifications
You must be signed in to change notification settings - Fork 10
/
pyproject.toml
123 lines (111 loc) · 2.98 KB
/
pyproject.toml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
[build-system]
# Should be mirrored in requirements-build.txt
requires = [
"cmake>=3.21",
"ninja",
"packaging",
"setuptools >= 49.4.0",
"torch == 2.3.0",
"wheel",
]
build-backend = "setuptools.build_meta"
[tool.ruff]
# Allow lines to be as long as 80.
line-length = 80
exclude = [
# External file, leaving license intact
"examples/fp8/quantizer/quantize.py"
]
[tool.ruff.lint]
select = [
# pycodestyle
"E",
# Pyflakes
"F",
# pyupgrade
# "UP",
# flake8-bugbear
"B",
# flake8-simplify
"SIM",
# isort
# "I",
"G",
]
ignore = [
# star imports
"F405", "F403",
# lambda expression assignment
"E731",
# Loop control variable not used within loop body
"B007",
]
[tool.mypy]
python_version = "3.8"
ignore_missing_imports = true
check_untyped_defs = true
follow_imports = "skip"
files = "vllm"
# TODO(woosuk): Include the code from Megatron and HuggingFace.
exclude = [
"vllm/model_executor/parallel_utils/|vllm/model_executor/models/",
# Ignore triton kernels in ops.
'vllm/attention/ops/.*\.py$'
]
[tool.codespell]
ignore-words-list = "dout, te, indicies, subtile"
skip = "./tests/prompts,./benchmarks/sonnet.txt,./tests/lora/data,./build"
[tool.isort]
use_parentheses = true
skip_gitignore = true
[tool.pytest.ini_options]
markers = [
"skip_global_cleanup",
"vlm: run tests for vision language models only",
]
[tool.coverage.run]
# collect context info for reporting, e.g. which tests exercised a line
dynamic_context = "test_function"
# do not examine any files matching these patterns
omit = [
# coverage for these devices is to be ignored until we plan to deploy them
"*neuron*",
"*cpu*",
# these model pathways are to be ignored for now
"*arctic*",
"*falcon*",
"*deepspeedfp*",
# we are not testing AMD devices at the moment
"**/rocm_flash_attn.py"
]
[tool.coverage.report]
# don't bother to report files w/out executable code, e.g. __init__.py
skip_empty = true
# coverage will not report on lines, or code blocks, matching these expressions
exclude_also = [
"def __repr__",
"raise AssertionError",
"raise NotImplementedError",
"if __name__ == .__main__.:",
"if TYPE_CHECKING:",
"@(abc\\.)?abstractmethod",
# coverage for these devices is to be ignored until we plan to deploy them
'engine_config.device_config.device_type == "cpu"',
'engine_config.device_config.device_type == "neuron"',
"elif is_neuron():",
"elif is_cpu():",
# customop.py has some placeholder code not yet executed
"return self.forward_cuda",
# ignore some legacy code that we won't support in nm-vllm
"enable_hf_transfer()",
"def convert_bin_to_safetensor_file",
"def _shared_pointers",
"def np_cache_weights_iterator",
"def convert_pyslice_to_tensor",
]
[tool.coverage.html]
# include the context info in the HTML report
show_contexts = true
[tool.coverage.json]
# include the context info in the JSON report
show_contexts = true