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 16, 15:45 UTC
[CI/Build] fix setuptools-scm usage (#8771)
Daniele committed
Sep 24, 2024 at 21:38 UTC
2467b642dd9bde32a334fe5967efd78a53aa49da
3 files changed
+7
-10
.gitignore
+2
-5
@@ -1,5 +1,5 @@
1
-# vllm commit id, generated by setup.py
2
-vllm/commit_id.py
1
+# version file generated by setuptools-scm
2
+/vllm/_version.py
3
4
# vllm-flash-attn built from source
5
vllm/vllm_flash_attn/
@@ -196,8 +196,5 @@ _build/
196
*_hip*
197
hip_compat.h
198
199
-# version file generated by setuptools-scm
200
-/vllm/_version.py
201
-
199
# Benchmark dataset
200
benchmarks/*.json
pyproject.toml
-3
@@ -51,9 +51,6 @@ ignore = [
51
"UP032",
52
]
53
54
-[tool.setuptools_scm]
55
-version_file = "vllm/_version.py"
56
-
54
[tool.mypy]
55
python_version = "3.8"
56
setup.py
+5
-2
@@ -354,12 +354,15 @@ def get_path(*filepath) -> str:
354
355
356
def get_vllm_version() -> str:
357
- version = get_version()
357
+ version = get_version(
358
+ write_to="vllm/_version.py", # TODO: move this to pyproject.toml
359
+ )
360
+
361
sep = "+" if "+" not in version else "." # dev versions might contain +
362
363
if _no_device():
364
if envs.VLLM_TARGET_DEVICE == "empty":
362
- version += "+empty"
365
+ version += f"{sep}empty"
366
elif _is_cuda():
367
cuda_version = str(get_nvcc_cuda_version())
368
if cuda_version != MAIN_CUDA_VERSION: