View source on GitHub

Class NoDependency

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 == "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).


View source


Initialize self. See help(type(self)) for accurate signature.