How to reduce space between mnist plots in python?

Question

I'm trying to write a code for a function that displays the first 20 misclassified images in the form of the output below.

Try 1

use subplots_adjust

#Try 1
DISPLAY_CNT = 20
FIG_ROW_CNT = 2
_FIG_COL_CNT = int(DISPLAY_CNT / FIG_ROW_CNT)
_TOTAL_DISPLAY_CNT = FIG_ROW_CNT * _FIG_COL_CNT
def display_first_20(list):
    j = 1
    fig = plt.figure(figsize=(15, 15))
    for i in range(_TOTAL_DISPLAY_CNT):
        plt.subplot(FIG_ROW_CNT, _FIG_COL_CNT, j)
        img = x_train[i]
        img = img.reshape(28, 28)
        plt.imshow(img)
        plt.title("{}/{}".format(classify(i), d_test[i]))
        plt.axis('off')
        j += 1

    plt.subplots_adjust(left=0.1, right=0.9, bottom=0.1, top=0.9,
                        hspace=0.1, wspace=0.1)
    plt.show()

fail_list = failed_list()
display_first_20(fail_list[:DISPLAY_CNT])

but there's big white blank between rows. and I also tried changing hspace & wspace,still doesn't work.

so I changed subplots_adjust to fig.tight_layout()

Try 2

use fig.tight_layout()

DISPLAY_CNT = 20
FIG_ROW_CNT = 2
_FIG_COL_CNT = int(DISPLAY_CNT / FIG_ROW_CNT)

def display_first_20(list):
    j = 1
    fig = plt.figure(figsize=(20, 10))
    for i in list:
        plt.subplot(FIG_ROW_CNT, _FIG_COL_CNT, j)
        img = x_train[i]
        img = img.reshape(28, 28)
        plt.imshow(img)
        plt.title("{}/{}".format(classify(i), d_test[i]))
        plt.axis('off')
        j += 1

    fig.tight_layout()
    plt.show()

fail_list = failed_list()
display_first_20(fail_list[:DISPLAY_CNT])

reduced some but still big to compare with first pic.

How can I make the code to display like first pic?