Skip to content

refactor: use the shared zero-variance check for native Tinker advantages #1122

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
bradhilton wants to merge 1 commit into
base: main
Choose a base branch
from
Open

refactor: use the shared zero-variance check for native Tinker advantages #1122

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
3 changes: 2 additions & 1 deletion src/art/tinker_native/data.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@
)
from ..trajectories._selection import ModelSelector, resolve_training_model
from ..types import MessagesAndChoices
from ..utils.grpo import rewards_have_zero_variance


def create_conversation_prefix_with_tools(
Expand All @@ -34,7 +35,7 @@ def compute_advantages(
) -> list[float]:
if not rewards:
return []
if all(abs(reward - rewards[0]) <= 1e-12 for reward in rewards[1:]):
if rewards_have_zero_variance(rewards):
return [0.0] * len(rewards)
rewards_tensor = torch.tensor(rewards, dtype=torch.float32)
centered = rewards_tensor - rewards_tensor.mean()
Expand Down
Loading