Adam optimizer iteration gives me float with step 0.5

Question

when i use this function to print optimizer iterations during training it gives me this output

epochs 1/100
step    adam iterations
0       1.000
1       1.500
2       2.000
3       2.500
....
99      50.500

epochs 2/100
step    adam iterations
0       101.000
1       101.500
2       102.000
3       102.500

and so on .

def get_lr_metric(optimizer):
    def lr(y_true, y_pred):
        return optimizer.iterations
    return lr


model.compile(,   metrics=[get_lr_metric(optimizer)])

i can not understand how adam optimizer iteration could be float with step equal t0 0.5

i expect the output to be like this

epochs 1/100
step    adam iterations
0       1
1       2
2       3
3       4
....
99      100

epochs 2/100
step    adam iterations
0       101
1       102
2       103
3       104