Sigmoid Activation

Apply the logistic sigmoid $1\mathrm{/}(1+e^{-x})1/(1+e^\{-x\})$ element-wise to an array of logits without overflowing for large positive or negative values.

Implement solve(x).

Returns. Return a float NumPy array of the same shape with values in $[0,1][0,1]$.

Examples

Example 1

Input:

solve([0.0])

Output:

[0.5]

Example 2

Input:

solve([2.0, -2.0])

Output:

[0.880797, 0.119203]

Example 3

Input:

solve([800.0, -800.0])

Output:

[1.0, 0.0]