visualization.py

import numpy as np
import matplotlib.pyplot as plt

def plot_kernels(tensor, number_cols=5, m_interpolation = 'bilinear'):
    '''
    Function to visualize the kernels.
    
    Arguments:
        tensor:
        number_cols: number of columns to be displayed
        m_interpolation: interpolation methods matplotlib. See in:
    
        https://matplotlib.org/gallery/images_contours_and_fields/interpolation_methods.html
    '''
    
    number_kernels = tensor.shape[0]
    number_rows = 1 + number_kernels // number_cols
    fig = plt.figure(figsize = (number_cols, number_rows))
    for i in range(number_kernels):
        ax1 = fig.add_subplot(number_rows, number_cols, i + 1)
        ax1.imshow(tensor[i][0, :, :], interpolation = m_interpolation, cmap='gray')
        ax1.axis('off')