CQM to BQM running on LeapHybridSampler and LeapHybridBQMSampler

I am successfully using CQM with LeapHybridCQMSampler for solving a problem (beautifully I might add).

I am also succesfully using dimod.cqm_to_bqm() to run the CQM model as a BQM using SimulatedAnnealingSampler for testing and benchmarks.

Now I would like to run the same BQM using LeapHybridBQMSampler and I am getting the following error:

TypeError: expected input vartype to be one of: Vartype.SPIN, 'SPIN', {-1, 1}, Vartype.BINARY, 'BINARY', or {0, 1}; received None.

What might I be missing to make the BQM from cqm_to_bqm compliant with LeapHybridBQMSampler ? Is it possible?



Here is the full stack

/usr/local/lib/python3.7/dist-packages/dwave/system/samplers/leap_hybrid_sampler.py in sample(self, bqm, time_limit, **kwargs)
    223         """
    224         if not isinstance(bqm, BQM):
--> 225             bqm = BQM(bqm)
    227         num_vars = bqm.num_variables

/usr/local/lib/python3.7/dist-packages/dimod/binary/binary_quadratic_model.py in __init__(self, offset, vartype, dtype, *args)
    167                 self._init_bqm(args[0], vartype=args[0].vartype, dtype=dtype)
    168             else:
--> 169                 self._init_empty(vartype=args[0], dtype=dtype)
    170         elif len(args) == 2:
    171             # BQM(bqm, vartype), BQM(n, vartype) or BQM(M, vartype)

/usr/local/lib/python3.7/dist-packages/dimod/binary/binary_quadratic_model.py in _init_empty(self, vartype, dtype)
    262     def _init_empty(self, vartype, dtype):
    263         dtype = self.DEFAULT_DTYPE if dtype is None else dtype
--> 264         self.data = type(self)._DATA_CLASSES[np.dtype(dtype)](vartype)
    266     def __init_subclass__(cls, default_dtype=np.float64, **kwargs):

/usr/local/lib/python3.7/dist-packages/dimod/binary/cybqm/cybqm_template.pyx.pxi in dimod.binary.cybqm.cybqm_float64.cyBQM_template.__init__()

/usr/local/lib/python3.7/dist-packages/dimod/binary/cybqm/cybqm_template.pyx.pxi in dimod.binary.cybqm.cybqm_float64.cyBQM_template.change_vartype()

/usr/local/lib/python3.7/dist-packages/dimod/vartypes.py in as_vartype(vartype, extended)
    171             raise _vartype_miss(vartype, extended) from None
    172     else:
--> 173         raise _vartype_miss(vartype, extended)
    175     if not extended and vartype != SPIN and vartype != BINARY:

TypeError: expected input vartype to be one of: Vartype.SPIN, 'SPIN', {-1, 1}, Vartype.BINARY, 'BINARY', or {0, 1}; received None.





  • Hello,

    It sounds like you might be using the object returned from the dimod.cqm_to_bqm() function directly.

    If you take a look at the reference docs, this function actually returns a tuple:

    In order to access the BQM returned, the function should be called like this:

    bqm, invert = dimod.cqm_to_bqm(cqm)

    Hopefully this resolves your issue. 
    Please feel free to reach out with more questions or for clarification.

    Comment actions Permalink
  • Hello David,

    Well... that was interesting. An additional condition I added to select Quantum BQM vs Classical was inadvertently excluding the cqm_to_bqm in the specific case of wanting to use CQM->BQM Simulated -> BQM->Quantum.

    Thanks for the pointer, I could not see the forest from the trees on this one.



    Comment actions Permalink
  • Hello,

    It sounds like you have sorted out the problem.
    Was this specific to your implementation or could we possibly make something more clear in the documentation?

    Thank you for reaching out to us.

    Comment actions Permalink
  • Hello,


    All good, this was entirely a coding oversight on my part.

    Thank you.

    Comment actions Permalink

Please sign in to leave a comment.

Didn't find what you were looking for?

New post