diff --git a/maths/softmax.py b/maths/softmax.py index ff872e2d17d0..2ebbc5eb8c46 100644 --- a/maths/softmax.py +++ b/maths/softmax.py @@ -33,15 +33,10 @@ def softmax(vector: np.ndarray, axis: int | None = -1) -> np.ndarray: probability distribution that sums to 1. Raises: - ValueError: If ``vector`` is empty or cannot be converted to a numeric - float array (for example a string or a dict). + ValueError: If ``vector`` is empty, cannot be converted to a numeric + float array, or contains non-finite values. numpy.exceptions.AxisError: If ``axis`` is out of bounds for the input. - Note: - If the input contains ``NaN`` or ``inf`` the result will contain - ``NaN`` along the affected axis; softmax is only meaningful for finite - real inputs. - The softmax vector adds up to one. We need to ceil to mitigate precision. >>> float(np.ceil(np.sum(softmax([1, 2, 3, 4])))) @@ -77,7 +72,7 @@ def softmax(vector: np.ndarray, axis: int | None = -1) -> np.ndarray: >>> float(np.round(np.sum(softmax(mat, axis=None)), 6)) 1.0 - Empty, non-numeric and out-of-bounds inputs raise clear errors: + Empty, non-numeric, non-finite and out-of-bounds inputs raise clear errors: >>> softmax([]) Traceback (most recent call last): @@ -91,6 +86,10 @@ def softmax(vector: np.ndarray, axis: int | None = -1) -> np.ndarray: Traceback (most recent call last): ... numpy.exceptions.AxisError: axis 3 is out of bounds for array of dimension 1 + >>> softmax([1.0, np.nan]) + Traceback (most recent call last): + ... + ValueError: softmax input must contain only finite values """ # Convert input to a float numpy array, turning numpy's terse conversion # errors into a clear message about the unsupported input type. @@ -100,9 +99,11 @@ def softmax(vector: np.ndarray, axis: int | None = -1) -> np.ndarray: error_message = f"softmax input must be numeric, got {type(vector).__name__}" raise ValueError(error_message) from exc - # Handle empty input + # Handle empty and non-finite input if vector.size == 0: raise ValueError("softmax input must be non-empty") + if not np.all(np.isfinite(vector)): + raise ValueError("softmax input must contain only finite values") # Validate axis (None means "treat the whole array as one distribution") if axis is not None: