Skip to content

Fix: warnings in softplus and inverse_softplus #270 #275

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
Dec 11, 2024
Merged

Fix: warnings in softplus and inverse_softplus #270 #275

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
8 changes: 6 additions & 2 deletions bayesflow/utils/numpy_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,9 @@ def inverse_shifted_softplus(

def inverse_softplus(x: np.ndarray, beta: float = 1.0, threshold: float = 20.0) -> np.ndarray:
"""Numerically stabilized inverse softplus function."""
return np.where(beta * x > threshold, x, np.log(beta * np.expm1(x)) / beta)
with np.errstate(over="ignore"):
expm1_x = np.expm1(x)
return np.where(beta * x > threshold, x, np.log(beta * expm1_x) / beta)


def one_hot(indices: np.ndarray, num_classes: int, dtype: str = "float32") -> np.ndarray:
Expand All @@ -37,4 +39,6 @@ def shifted_softplus(

def softplus(x: np.ndarray, beta: float = 1.0, threshold: float = 20.0) -> np.ndarray:
"""Numerically stabilized softplus function."""
return np.where(beta * x > threshold, x, np.log1p(np.exp(beta * x)) / beta)
with np.errstate(over="ignore"):
exp_beta_x = np.exp(beta * x)
return np.where(beta * x > threshold, x, np.log1p(exp_beta_x) / beta)
Loading