Warning message: Distribution subclass TanhTransformedDistribution inherits _parameter_properties from its parent ...

I can see a bunch of super-long warning messages generated from Distribution classes:

https://github.com/tensorflow/probability/blob/main/tensorflow_probability/python/distributions/distribution.py#L291-L325

W1103 21:19:41.807866 140018421924672 distribution.py:296]                                                                        
Distribution subclass TanhTransformedDistribution inherits `_parameter_properties from its parent (TransformedDistribution)       
while also redefining `__init__`. The inherited annotations cover the following                                                   
parameters: dict_keys(['distribution', 'bijector']). It is likely that these do not match the subclass parameters.                
This may lead to errors when computing batch shapes, slicing into batch                                                           
dimensions, calling `.copy()`, flattening the distribution as a CompositeTensor                                                   
(e.g., when it is passed or returned from a `tf.function`), and possibly other                                                    
cases. The recommended pattern for distribution subclasses is to define a new                                                     
`_parameter_properties` method with the subclass parameters, and to store the                                                     
corresponding parameter values as `self._parameters` in `__init__`, after                                                         
calling the superclass constructor:                                                                                               
                                                                                                                                  
class MySubclass(tfd.SomeDistribution):                                                                                           
                                                                                                                                  
  def __init__(self, param_a, param_b):                                                                                           
    parameters = dict(locals())                                                                                                   
    # ... do subclass initialization ...                                                                                          
    super(MySubclass, self).__init__(**base_class_params)                                                                         
    # Ensure that the subclass (not base class) parameters are stored.                                                            
    self._parameters = parameters                                                                                                 
                                                                                                                                  
  def _parameter_properties(self, dtype, num_classes=None):                                                                       
    return dict(                                                                                                                  
      # Annotations may optionally specify properties, such as `event_ndims`,                                                     
      # `default_constraining_bijector_fn`, `specifies_shape`, etc.; see                                                          
      # the `ParameterProperties` documentation for details.                                                                      
      param_a=tfp.util.ParameterProperties(),                                                                                     
      param_b=tfp.util.ParameterProperties())                                                                                                                                                                                                                       

Caller: Just a plain normal TanhTransformedDistribution(distribution): see
https://github.com/deepmind/acme/blob/4544216e322b849e07ce1d964d2129054e611568/acme/jax/networks/distributional.py#L230-L231

I was simply running dm-acme example code. How to reproduce: please run https://github.com/deepmind/acme/blob/master/examples/gym/run_sac.py

The packages I used (non-exhaustive list):

tf-nightly==2.8.0.dev20211102
tfp-nightly==0.15.0.dev20211103
dm-acme (nightly, HEAD)
dm-haiku==0.0.5

Questions:

• Why is this happening? Is this a bug or invalid usage (perhaps happened due to dm-haiku)? From the user code and implementation of NormalTanhDistribution, I don't see anything suspicious or wrong.
• Is it possible to disable/ignore such warnings? The package is using a root logger, so module-wise disabling seems not easy.

Start with tensorflow_probability/python/distributions/distribution.py#L291-L325 and reproduce the warning using dm-acme's examples/gym/run_sac.py. Compare the warning's inherited parameters with TanhTransformedDistribution and determine whether the behavior or warning handling is incorrect; done means the reported warning is appropriately resolved or its suppression behavior is clarified.