from collections import namedtuple
import glob
from iris import log
# Total guess based on the rule of thumb from this question:
# http://stackoverflow.com/questions/10565868/multi-layer-perceptron-mlp-architecture-criteria-for-choosing-number-of-hidde
NUM_HIDDEN = 21
# Sepal Length, Sepal Width, Petal Length and Petal Width == 4 features.
NUM_FEATURES = 4
# Iris-setosa, Iris-versicolor and Iris-virginica == 3
NUM_LABELS = 3
_logger = log.get_logger()
# Data structure used to keep track of important parameters used in training our
# model.
Topology = namedtuple("Topology", "x y t w_hidden b_hidden hidden w_out b_out")
def generate_weight(shape, name):
"""
TF variable filled with random values which follow a normal distribution.
Parameters
----------
shape : 1-D Tensor or Array
Corresponds to the shape parameter of tf.random_normal.
name : str
from collections import namedtuple
import glob
from iris import log
# Total guess based on the rule of thumb from this question:
# http://stackoverflow.com/questions/10565868/multi-layer-perceptron-mlp-architecture-criteria-for-choosing-number-of-hidde
NUM_HIDDEN = 21
# Sepal Length, Sepal Width, Petal Length and Petal Width == 4 features.
NUM_FEATURES = 4
# Iris-setosa, Iris-versicolor and Iris-virginica == 3
NUM_LABELS = 3
_logger = log.get_logger()
# Data structure used to keep track of important parameters used in training our
# model.
Topology = namedtuple(
"Topology",
"x y t w_hidden b_hidden hidden w_out b_out")
def generate_weight(shape, name):
"""
TF variable filled with random values which follow a normal distribution.
Parameters
----------
shape : 1-D Tensor or Array