We know tuple can represent the same collection, for example, two-dimensional coordinates of a point can be expressed as:
p = (1, 2)
However, see (1, 2), it is difficult to see this tuple is used to represent a coordinate. At this time, namedtuple came in handy.
usage:
Examples of the use of a coordinate namedtuple represented as follows:
from collections import namedtuple
Point = namedtuple('Point', ['x', 'y'])
p = Point(1, 2)
print(p.x,p.y)
Output: `2 '
Another example of Densenet:
from collections import namedtuple
DensenetParams = namedtuple('DensenetParameters', ['num_classes',
'first_output_features',
'layers_per_block',
'growth_rate',
'bc_mode',
'is_training',
'dropout_keep_prob'
])
default_params = DensenetParams(
num_classes = 10,
first_output_features = 24,
layers_per_block = 12,
growth_rate = 12,
bc_mode = False,
is_training = True,
dropout_keep_prob = 0.8,
)
print(default_params.num_classes)
The output is:10