tf.contrib.checkpoint.NoDependency

View source on GitHub

Allows attribute assignment to Trackable objects with no dependency.

Example usage:

obj = Trackable()
obj.has_dependency = tf.Variable(0., name="dep")
obj.no_dependency = NoDependency(tf.Variable(1., name="nodep"))
assert obj.no_dependency.name == "nodep:0"

obj in this example has a dependency on the variable "dep", and both attributes contain un-wrapped Variable objects.

NoDependency also works with tf.keras.Model, but only for checkpoint dependencies: wrapping a Layer in NoDependency will assign the (unwrapped) Layer to the attribute without a checkpoint dependency, but the Model will still track the Layer (so it will appear in Model.layers, and its variables will appear in Model.variables).