tf.placeholder inside a class in TF 2.0 - tensorflow

Question

I'm trying to change a code I have written in TF 1.0 to TF 2.0 and I'm having difficulties with replacing the tf.placeholder inside a class function. My code is the following
class User:
x = tf.placeholder(tf.float32,shape=[None,784])
y_true = tf.placeholder(tf.float32, [None, 10])
W1 = tf.Variable(tf.random.truncated_normal([7840,1], stddev=0.1))
lambda_W = tf.Variable(tf.zeros([7840,1]))
W = tf.reshape(W1,[784, 10])
ylogits = W*x
y = tf.nn.softmax(ylogits)
def __init__(self):
pass
Is there a way to replace tf.placeholder inside the class to make code running in TF 2.0?

Answer 1

Firstly, I think you intended to have each of those objects created per instance of the class, not one for the whole class as it is now. I also think your product between W and x was meant to be a matrix product, not an element-wise product, which would not work with the given shapes:
class User:
def __init__(self):
self.x = tf.placeholder(tf.float32,shape=[None,784])
self.y_true = tf.placeholder(tf.float32, [None, 10])
self.W1 = tf.Variable(tf.random.truncated_normal([7840,1], stddev=0.1))
self.lambda_W = tf.Variable(tf.zeros([7840,1]))
self.W = tf.reshape(W1,[784, 10])
self.ylogits = self.x # self.W
self.y = tf.nn.softmax(ylogits)
To use it in TensorFlow 2.x, you would remove the placeholders and simply perform the operations each time with every new input, for example with a new function call:
class User:
def __init__(self):
self.W1 = tf.Variable(tf.random.truncated_normal([7840,1], stddev=0.1))
self.lambda_W = tf.Variable(tf.zeros([7840,1]))
self.W = tf.reshape(W1,[784, 10])
def call(self, x):
ylogits = self.x # self.W
return tf.nn.softmax(ylogits)
You could use this as:
user1 = User()
x = ... # Get some data
y = user1.call(x)
Or if you like to be more "idiomatic", you could use __call__:
class User:
def __init__(self):
self.W1 = tf.Variable(tf.random.truncated_normal([7840,1], stddev=0.1))
self.lambda_W = tf.Variable(tf.zeros([7840,1]))
self.W = tf.reshape(W1,[784, 10])
def __call__(self, x):
ylogits = x # W
return tf.nn.softmax(ylogits)
And then you would do:
user1 = User()
x = ... # Get some data
y = user1(x)