Binary Cross-Entropy
Beginner
Compute the mean binary cross-entropy from real-valued logits and 0/1 labels using a numerically stable formula that never evaluates the sigmoid directly.
Implement solve(logits, labels).
Returns. Return a non-negative Python float. Very large logits must not overflow.
Examples
Example 1
Input:
solve([0.0, 0.0], [1.0, 0.0])
Output:
0.693147
Example 2
Input:
solve([2.0, -2.0, 1000.0], [1.0, 0.0, 1.0])
Output:
0.084619
Keywords
problemsetclassical-mllogistic-regression
Python
Run your code to see results
No results yet
Run your code to check it against the examples, or submit to run every test.