AttributeError: damp object has no attribute '_data_buffer'

[jazzskier]

I was trying to run the tutorial 1 about modeling, and got errors at:

model = Model(vp=v, origin=origin, shape=shape, spacing=spacing, space_order=2, nbl=10)

Could anyone help take a look? Thanks!

/content/devito/examples/seismic/model.py in __init__(self, origin, spacing, shape, space_order, vp, nbl, dtype, epsilon, delta, theta, phi, subdomains, **kwargs)
    201                  subdomains=(), **kwargs):
    202         super(Model, self).__init__(origin, spacing, shape, space_order, nbl, dtype,
--> 203                                     subdomains)
    204 
    205         # Create square slowness of the wave as symbol `m`

/content/devito/examples/seismic/model.py in __init__(self, origin, spacing, shape, space_order, nbl, dtype, subdomains, damp_mask)
     87             # Create dampening field as symbol `damp`
     88             self.damp = Function(name="damp", grid=self.grid)
---> 89             initialize_damp(self.damp, self.nbl, self.spacing, mask=damp_mask)
     90             self._physical_parameters = ['damp']
     91         else:

/content/devito/examples/seismic/model.py in initialize_damp(damp, nbl, spacing, mask)
     48 
     49     # TODO: Figure out why yask doesn't like it with dse/dle
---> 50     Operator(eqs, name='initdamp', dse='noop', dle='noop')()
     51 
     52 

/content/devito/devito/operator/operator.py in __call__(self, **kwargs)
    495 
    496     def __call__(self, **kwargs):
--> 497         self.apply(**kwargs)
    498 
    499     def apply(self, **kwargs):

/content/devito/devito/operator/operator.py in apply(self, **kwargs)
    561         # Build the arguments list to invoke the kernel function
    562         with self._profiler.timer_on('arguments'):
--> 563             args = self.arguments(**kwargs)
    564 
    565         # Invoke kernel function with args

/content/devito/devito/operator/operator.py in arguments(self, **kwargs)
    443     def arguments(self, **kwargs):
    444         """Arguments to run the Operator."""
--> 445         args = self._prepare_arguments(**kwargs)
    446         # Check all arguments are present
    447         for p in self.parameters:

/content/devito/devito/operator/operator.py in _prepare_arguments(self, **kwargs)
    353                 # E.g., SubFunctions
    354                 continue
--> 355             for k, v in p._arg_values(**kwargs).items():
    356                 if k in args and args[k] != v:
    357                     raise ValueError("Default `%s` is incompatible with other args as "

/content/devito/devito/types/dense.py in _arg_values(self, **kwargs)
    782                     values.update(self.grid._arg_defaults())
    783         else:
--> 784             values = self._arg_defaults(alias=self).reduce_all()
    785 
    786         return values

/content/devito/devito/types/dense.py in _arg_defaults(self, alias)
    742         """
    743         key = alias or self
--> 744         args = ReducerMap({key.name: self._data_buffer})
    745 
    746         # Collect default dimension arguments from all indices

/content/devito/devito/finite_differences/differentiable.py in __getattr__(self, name)
    128         if name in self._fd:
    129             return self._fd[name][0](self)
--> 130         raise AttributeError("%r object has no attribute %r" % (self.__class__, name))
    131 
    132     # Override SymPy arithmetic operators

AttributeError: damp object has no attribute '_data_buffer'

[FabioLuporini]

can you edit the original message and put the stack trace within a code block:

code ...

otherwise formatting and layout are messed up

now a question: what did you change to get this error? did it work the first time you run it?

[jazzskier]

Sure, I will edit the original message.

I am running the tutorial 1 on google colab notebooks. I followed the quickstart steps (conda virtual env and pip install) and created the environment. Nothing else is added. Then it failed with the error of attribute '_data_buffer'.

[image: image.png]

Thanks, Zhenzhen

On Thu, Jan 9, 2020 at 3:10 AM Fabio Luporini [email protected] wrote:

can you edit the original message and put the stack trace within a code block:

code ...

otherwise formatting and layout are messed up

now a question: what did you change to get this error? did it work the first time you run it?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/devitocodes/devito/issues/1047?email_source=notifications&email_token=AFUEET2B7KXN6QDQWP4CVJ3Q43SY7A5CNFSM4KERT2J2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEIPRVWQ#issuecomment-572463834, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFUEETZZ44IYBZQW332CVJ3Q43SY7ANCNFSM4KERT2JQ .

[FabioLuporini]

That's really odd. Will look into this.

Which cell is failing exactly?

[ggorman]

Are there any further updates on this?

[navjotk]

I have seen this error myself, recently. Forgot how I resolved it.

What python version are/were you using @jazzskier ?