Skip to content

Fix torch.load of transformers patch when enabling low_cpu_mem_usage #10

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.

Sign up for GitHub

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

Open
lausannel wants to merge 1 commit into
base: main
Choose a base branch
from
+5 −2
Open

Fix torch.load of transformers patch when enabling low_cpu_mem_usage #10

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
7 changes: 5 additions & 2 deletions flashmodels/patch/patch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,8 +19,11 @@ def rewrite_load():
"""Rewrite `torch.load` in `from_pretrain` in case to use mmap to reduce the CPU
memory pressure of loading multiple copies of data under multiple processes"""
source = inspect.getsource(transformers.modeling_utils)
modified = re.sub(r"torch\.load\((?![^)]*mmap[^)]*\))([^)]*)\)",
r"torch.load(\g<1>, mmap=True)", source)
modified = re.sub(
r"torch\.load\((?![^)]*mmap[^)]*\))([^)]*)\)",
r"torch.load(\g<1>, mmap=True)"
if version.parse(transformers.__version__) <= version.parse('4.36')
else r"torch.load(\g<1> mmap=True)", source)
modified = re.sub(r"partial\(torch.load,(?![^)]*mmap[^)]*\))([^)]*)\)",
r"partial(torch.load,\g<1>, mmap=True)", modified)
if (int(os.environ.get("LOCAL_RANK", 0)) == 0):
Expand Down