I have the following issue when running the code below:

import tensorflow.compat.v1 as tfc

image=tfc.Variable(tf.zeros((1,224,224,3)))

#It declares the image
x = tfc.placeholder(tfc.float32, (1,224, 224, 3))

#our trainable adversarial input
x_hat = image 

#Update ref by assigning value to it.
assign_op = tfc.assign(x_hat, x)

#Define learning rate as a placeholder
learning_rate = tfc.placeholder(tf.float32, ())

#defines y_hat as an int32 value
y_hat = tfc.placeholder(tf.int32, ())

#defines labels as one hot value with 1000 possible 0-1 values
labels = tfc.one_hot(y_hat, 12)

#loss used: Measures the probability error in discrete classification tasks in which the classes are mutually  
#exclusive (each entry is in exactly one class). 
#For example, each CIFAR-10 image is labeled with one and only one label: an image can be a dog or a truck, 
#but not both.
loss = tfc.nn.softmax_cross_entropy_with_logits(logits=logits, labels=[labels])

optim_step = tfc.train.GradientDescentOptimizer(learning_rate).minimize(loss, var_list=[x_hat])

Whenever I run this code I have the following error:

    ---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-41-b55e39746a98> in <module>
     17 #assign_op = tf.assign(x_hat, x)
     18 #The way the loss is optimized
---> 19 optim_step = tfc.train.GradientDescentOptimizer(learning_rate).minimize(loss, var_list=[x_hat])

/usr/local/lib/python3.8/dist-packages/tensorflow/python/training/optimizer.py in minimize(self, loss, global_step, var_list, gate_gradients, aggregation_method, colocate_gradients_with_ops, name, grad_loss)
    405     vars_with_grad = [v for g, v in grads_and_vars if g is not None]
    406     if not vars_with_grad:
--> 407       raise ValueError(
    408           "No gradients provided for any variable, check your graph for ops"
    409           " that do not support gradients, between variables %s and loss %s." %

ValueError: No gradients provided for any variable, check your graph for ops that do not support gradients, between variables ["<tf.Variable 'Variable_2:0' shape=(1, 224, 224, 3) dtype=float32>"] and loss Tensor("softmax_cross_entropy_with_logits_sg_11/Reshape_2:0", shape=(1,), dtype=float32).

What is wrong with my code?