Error in Black Box Likelihood Function Example with PyMC

Questions v5
1

Hello,

I’m trying out the Using a “black box” likelihood function example, but I’m encountering an error and it’s not working as expected.

Specifically, the error occurs with the following code:

grad_model.compile_dlogp()(ip)

The error message is as follows:

---------------------------------------------------------------------------
NotImplementedError                       Traceback (most recent call last)
Cell In[24], line 1
----> 1 grad_model.compile_dlogp()(ip)

File d:\anaconda\envs\pymc\Lib\site-packages\pymc\model\core.py:620, in Model.compile_dlogp(self, vars, jacobian, **compile_kwargs)
    604 def compile_dlogp(
    605     self,
    606     vars: Variable | Sequence[Variable] | None = None,
    607     jacobian: bool = True,
    608     **compile_kwargs,
    609 ) -> PointFunc:
    610     """Compiled log probability density gradient function.
    611 
    612     Parameters
   (...)
    618         Whether to include jacobian terms in logprob graph. Defaults to True.
    619     """
--> 620     return self.compile_fn(self.dlogp(vars=vars, jacobian=jacobian), **compile_kwargs)

File d:\anaconda\envs\pymc\Lib\site-packages\pymc\model\core.py:760, in Model.dlogp(self, vars, jacobian)
    758 cost = self.logp(jacobian=jacobian)
    759 cost = rewrite_pregrad(cost)
--> 760 return gradient(cost, value_vars)
...
---> 32     raise NotImplementedError("Gradient only implemented for scalar m and c")
     34 grad_wrt_m, grad_wrt_c = loglikegrad_op(m, c, sigma, x, data)
     36 # out_grad is a tensor of gradients of the Op outputs wrt to the function cost

NotImplementedError: Gradient only implemented for scalar m and c

Environment details:

  • python : 3.12.8
  • pytensor : 2.26.4
  • pymc : 5.19.1
  • arviz : 0.20.0
  • matplotlib: 3.10.0
  • scipy : 1.14.1
  • numpy : 1.26.4

I would appreciate it if you could help me understand the cause of this error and how to resolve it.

2

Hello, I have met same error as yours.
I tried to construct my own likelihood function with gradient refered to [Using a “black box” likelihood function]. But it failed in the “model.compile_dlogp()(ip)”, occuring same NotImplementedError.Then, I tried to run the example codes in colab. It turns out the codes in example doesn’t work :sweat_smile:
After a lot of investigations, I figure it out what happen !!!
First, if you delete the sentence in grad():" if m.type.dim!=0 ……", you would meet an error: “LokLIkeWithGrad.grad returned a term with 0 dimensions, but 1 are required”, and
the above sentence: “if hasattr(var, “ndim”) and term.ndim != var.ndim: raise ValueError(……）”
These errors mean that the first term returned from grad was a scalar, but the variable from the node is a vector with shape(1,).
The problem is clarified! And the solution is quite simple:
In make_node() of LogLikeWithGrad(Op), after converting inputs to tensor variables “m=pt.as_tensor(m)”, we could add a sentence “m = m[0]”, that could convert a vector with shape(1,) into a scalar with same dimensioin as the output of grad(). Besides, we could add some dimension judgment statement (such as "if m.type.ndim!=0 ") to avoid potential error.
Hope my experience could help you.

3

Thanks for investigating, I’ll take a look

4

Any progress figuring this out? I’m running into the same problem myself. Introducing a line m = m[0] after the tensor conversion m = pt.as_tensor(m) (and doing the same for c) doesn’t seem to fix the problem.

For what it’s worth, the older example available in the version history does seem to work properly, inasmuch as PyMC uses the NUTS sampler when the code is run. The code in the current example leads to the Slice sampler being invoked instead, presumably because the gradient function is not evaluating correctly.

(I’m entirely a newbie at this stuff, so my assessment of the situation may be way off. Feel free to disregard my ideas about what’s going wrong or how to fix it.)

5

Forgot to include my environment details:

  • python : 3.13.3
  • pytensor : 2.30.3
  • pymc : 5.22.0
  • arviz : 0.21.0
  • matplotlib-base: 3.10.3
  • scipy : 1.15.2
  • numpy : 2.2.6