Skip to content

use base version for version comparison #19587

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Jun 13, 2025
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion vllm/compilation/backends.py
Original file line number Diff line number Diff line change
Expand Up @@ -31,7 +31,7 @@
def make_compiler(compilation_config: CompilationConfig) -> CompilerInterface:
if compilation_config.use_inductor:
if envs.VLLM_USE_STANDALONE_COMPILE and is_torch_equal_or_newer(
"2.8.0"):
"2.8.0a"):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sorry, what does "a" in "2.8.0a" mean here?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

'a' is a standard version for alpha: https://peps.python.org/pep-0440/
current pytorch nightly has the version 2.8.0a0+git093fd47

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I see, thanks for the explanation. So the version parser thinks 2.8.0 > 2.8.0a then.

logger.debug("Using InductorStandaloneAdaptor")
return InductorStandaloneAdaptor()
else:
Expand Down
4 changes: 2 additions & 2 deletions vllm/model_executor/layers/quantization/torchao.py
Original file line number Diff line number Diff line change
Expand Up @@ -44,14 +44,14 @@ def __init__(self,
"""
# TorchAO quantization relies on tensor subclasses. In order,
# to enable proper caching this needs standalone compile
if is_torch_equal_or_newer("2.8.0"):
if is_torch_equal_or_newer("2.8.0a"):
os.environ["VLLM_TEST_STANDALONE_COMPILE"] = "1"
logger.info(
"Using TorchAO: Setting VLLM_TEST_STANDALONE_COMPILE=1")

# TODO: remove after the torch dependency is updated to 2.8
if is_torch_equal_or_newer(
"2.7.0") and not is_torch_equal_or_newer("2.8.0"):
"2.7.0") and not is_torch_equal_or_newer("2.8.0a"):
os.environ["VLLM_DISABLE_COMPILE_CACHE"] = "1"
logger.info("Using TorchAO: Setting VLLM_DISABLE_COMPILE_CACHE=1")
"""
Expand Down