Skip to content

Use implicit strategy as fallback when no strategy is available #72

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

Merged
merged 1 commit into from
Aug 4, 2025
+3 −2
Merged

Use implicit strategy as fallback when no strategy is available #72

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
5 changes: 3 additions & 2 deletions autoparallel/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@
from torch.distributed.tensor.placement_types import Replicate
from torch.utils._pytree import tree_flatten, tree_map_only

from .dtensor_util import get_op_strategy
from .dtensor_util import get_op_strategy, with_implicit_strategies
from .propagation_rules import (
TENSOR_FACTORY_OPS,
_op_partial_rules,
Expand Down Expand Up @@ -163,7 +163,8 @@ def get_placement_options(mesh, op, specs, user_args, user_kwargs):
if op in _op_partial_rules:
out_strat = _op_partial_rules[op](mesh, op_schema)
else:
out_strat = get_op_strategy(op, op_schema)
with with_implicit_strategies():
out_strat = get_op_strategy(op, op_schema)

propagate_tensor_meta(op, user_args, user_kwargs, out_strat)
fill_missing_redistribute_cost(op, specs, out_strat)
Expand Down
Loading