Attention Visualization
In this problem, you will be implementing a function to visualize attention weights from a given attention matrix. Attention matrices are used in transformer models to weight the importance of different input elements. Your function should take in an attention matrix and return a heatmap representation of the attention weights.
Example:
Suppose we have an attention matrix [[0.1, 0.3], [0.7, 0.2]]. The function should return a heatmap with the attention weights plotted.
Constraints:
* The input attention matrix should be a 2D numpy array.
* The heatmap should be a 2D numpy array with the same shape as the input attention matrix.
Test Cases
Test Case 1
Input:
[[0.1, 0.3], [0.7, 0.2]]Expected:
Heatmap of attention weightsTest Case 2
Input:
[[0.5, 0.5], [0.5, 0.5]]Expected:
Heatmap of attention weights+ 3 hidden test cases