# Value Error in pytensor.function

**URL:** <https://discourse.pymc.io/t/value-error-in-pytensor-function/16181>\
**Category:** version agnostic\
**Tags:** pytensor\
**Created:** [November 25, 2024, 5:18pm UTC](https://discourse.pymc.io/t/value-error-in-pytensor-function/16181 "2024-11-25T17:18:34Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![chenxiating](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/chenxiating/32/8769_2.png) [@chenxiating](https://discourse.pymc.io/u/chenxiating)\
**Post date:** [November 25, 2024, 5:18pm UTC](https://discourse.pymc.io/t/value-error-in-pytensor-function/16181/1 "2024-11-25T17:18:34Z")

</div>

New to PyMC - I’m working through the tutorial on reinforced learning. My package versions are: Python (3.8.5), PyMC (5.6.1), PyTensor (2.12.3), Numpy (1.24.4), Arviz (0.15.1), Cython (0.29.33).I am getting a Value Error while running `pytensor_llik_td = pytensor.function( inputs=[alpha, beta], outputs=neg_loglike, on_unused_input="ignore" )` in section [“Estimating the learning parameter via PyMC”](https://www.pymc.io/projects/examples/en/latest/case_studies/reinforcement_learning.html#estimating-the-learning-parameters-via-pymc). Can someone help me figure out what to do?

Specifically, here are my errors:

```py
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/link/vm.py:1243, in VMLinker.make_all(self, profiler, input_storage, output_storage, storage_map)
   1239 # no-recycling is done at each VM. __call__ So there is
   1240 # no need to cause duplicate c code by passing
   1241 # no_recycling here.
   1242 thunks.append(
-> 1243 node.op.make_thunk(node, storage_map, compute_map, [], impl=impl)
   1244 )
   1245 linker_make_thunk_time[node] = time.perf_counter() - thunk_start

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/scan/op.py:1512, in Scan.make_thunk(self, node, storage_map, compute_map, no_recycling, impl)
   1510 raise MissingGXX
-> 1512 from . import scan_perform_ext
   1514 cython_mintaps = np.asarray(self.mintaps, dtype="int32")

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/scan/scan_perform_ext.py:8
      1 """
      2 
      3 To update the `Scan` Cython code you must
   (...)
      6 
      7 """
----> 8 from pytensor.scan.scan_perform import get_version, perform # noqa: F401, E402
     11 version = 0.326 # must match constant returned in function get_version()

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/scan/scan_perform.pyx:1, in init pytensor.scan.scan_perform()

ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject

During handling of the above exception, another exception occurred:

ValueError Traceback (most recent call last)
Cell In[31], line 36
     33 logp_actions = logp_actions[pt.arange(actions_.shape[0] - 1), actions_[1:]]
     34 neg_loglike = -pt.sum(logp_actions)
---> 36 pytensor_llik_td = pytensor.function(
     37 inputs=[alpha, beta], outputs=neg_loglike, on_unused_input="ignore"
     38 )
     39 result = pytensor_llik_td(true_alpha, true_beta)
     40 float(result)

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/compile/function/ __init__.py:315, in function(inputs, outputs, mode, updates, givens, no_default_updates, accept_inplace, name, rebuild_strict, allow_input_downcast, profile, on_unused_input)
    309 fn = orig_function(
    310 inputs, outputs, mode=mode, accept_inplace=accept_inplace, name=name
    311 )
    312 else:
    313 # note: pfunc will also call orig_function -- orig_function is
    314 # a choke point that all compilation must pass through
--> 315 fn = pfunc(
    316 params=inputs,
    317 outputs=outputs,
    318 mode=mode,
    319 updates=updates,
    320 givens=givens,
    321 no_default_updates=no_default_updates,
    322 accept_inplace=accept_inplace,
    323 name=name,
    324 rebuild_strict=rebuild_strict,
    325 allow_input_downcast=allow_input_downcast,
    326 on_unused_input=on_unused_input,
    327 profile=profile,
    328 output_keys=output_keys,
    329 )
    330 return fn

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/compile/function/pfunc.py:367, in pfunc(params, outputs, mode, updates, givens, no_default_updates, accept_inplace, name, rebuild_strict, allow_input_downcast, profile, on_unused_input, output_keys, fgraph)
    353 profile = ProfileStats(message=profile)
    355 inputs, cloned_outputs = construct_pfunc_ins_and_outs(
    356 params,
    357 outputs,
   (...)
    364 fgraph=fgraph,
    365 )
--> 367 return orig_function(
    368 inputs,
    369 cloned_outputs,
    370 mode,
    371 accept_inplace=accept_inplace,
    372 name=name,
    373 profile=profile,
    374 on_unused_input=on_unused_input,
    375 output_keys=output_keys,
    376 fgraph=fgraph,
    377 )

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/compile/function/types.py:1756, in orig_function(inputs, outputs, mode, accept_inplace, name, profile, on_unused_input, output_keys, fgraph)
   1744 m = Maker(
   1745 inputs,
   1746 outputs,
   (...)
   1753 fgraph=fgraph,
   1754 )
   1755 with config.change_flags(compute_test_value="off"):
-> 1756 fn = m.create(defaults)
   1757 finally:
   1758 t2 = time.perf_counter()

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/compile/function/types.py:1649, in FunctionMaker.create(self, input_storage, storage_map)
   1646 start_import_time = pytensor.link.c.cmodule.import_time
   1648 with config.change_flags(traceback __limit=config.traceback__ compile_limit):
-> 1649 _fn, _i, _o = self.linker.make_thunk(
   1650 input_storage=input_storage_lists, storage_map=storage_map
   1651 )
   1653 end_linker = time.perf_counter()
   1655 linker_time = end_linker - start_linker

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/link/basic.py:254, in LocalLinker.make_thunk(self, input_storage, output_storage, storage_map, **kwargs)
    247 def make_thunk(
    248 self,
    249 input_storage: Optional["InputStorageType"] = None,
   (...)
    252 **kwargs,
    253 ) -> Tuple["BasicThunkType", "InputStorageType", "OutputStorageType"]:
--> 254 return self.make_all(
    255 input_storage=input_storage,
    256 output_storage=output_storage,
    257 storage_map=storage_map,
    258 )[:3]

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/link/vm.py:1252, in VMLinker.make_all(self, profiler, input_storage, output_storage, storage_map)
   1250 thunks[-1].lazy = False
   1251 except Exception:
-> 1252 raise_with_op(fgraph, node)
   1254 t1 = time.perf_counter()
   1256 if self.profile:

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/link/utils.py:535, in raise_with_op(fgraph, node, thunk, exc_info, storage_map)
    530 warnings.warn(
    531 f"{exc_type} error does not allow us to add an extra error message"
    532 )
    533 # Some exception need extra parameter in inputs. So forget the
    534 # extra long error message in that case.
--> 535 raise exc_value.with_traceback(exc_trace)

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/link/vm.py:1243, in VMLinker.make_all(self, profiler, input_storage, output_storage, storage_map)
   1238 thunk_start = time.perf_counter()
   1239 # no-recycling is done at each VM. __call__ So there is
   1240 # no need to cause duplicate c code by passing
   1241 # no_recycling here.
   1242 thunks.append(
-> 1243 node.op.make_thunk(node, storage_map, compute_map, [], impl=impl)
   1244 )
   1245 linker_make_thunk_time[node] = time.perf_counter() - thunk_start
   1246 if not hasattr(thunks[-1], "lazy"):
   1247 # We don't want all ops maker to think about lazy Ops.
   1248 # So if they didn't specify that its lazy or not, it isn't.
   1249 # If this member isn't present, it will crash later.

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/scan/op.py:1512, in Scan.make_thunk(self, node, storage_map, compute_map, no_recycling, impl)
   1509 if impl == "py":
   1510 raise MissingGXX
-> 1512 from . import scan_perform_ext
   1514 cython_mintaps = np.asarray(self.mintaps, dtype="int32")
   1516 n_outs = self.info.n_mit_mot + self.info.n_mit_sot + self.info.n_sit_sot

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/scan/scan_perform_ext.py:8
      1 """
      2 
      3 To update the `Scan` Cython code you must
   (...)
      6 
      7 """
----> 8 from pytensor.scan.scan_perform import get_version, perform # noqa: F401, E402
     11 version = 0.326 # must match constant returned in function get_version()
     12 assert version == get_version(), (
     13 "Invalid extension, check the installation process, "
     14 "could be problem with .pyx file or Cython ext build process."
     15 )

File ~/.pyenv/versions/3.8.5/lib/python3.8/site-packages/pytensor/scan/scan_perform.pyx:1, in init pytensor.scan.scan_perform()

ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject
Apply node that caused the error: Scan{scan_fn, while_loop=False, inplace=all}(150, [1 0 0 0 0 ... 1 1
 1 1], [0 1 1 1 0 ... 1 0
 1 1], SetSubtensor{:stop}.0, alpha)
Toposort index: 3
Inputs types: [TensorType(int64, shape=()), TensorType(int32, shape=(150,)), TensorType(int32, shape=(150,)), TensorType(float64, shape=(150, 2)), TensorType(float64, shape=())]

HINT: Use a linker other than the C linker to print the inputs' shapes and strides.
HINT: Re-running with most PyTensor optimizations disabled could provide a back-trace showing when this node was created. This can be done by setting the PyTensor flag 'optimizer=fast_compile'. If that does not work, PyTensor optimizations can be disabled with 'optimizer=None'.
HINT: Use the PyTensor flag `exception_verbosity=high` for a debug print-out and storage map footprint of this Apply node.

```

---

<div class="post-metadata">

**Author:** ![jessegrabowski](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/jessegrabowski/32/5010_2.png) [@jessegrabowski](https://discourse.pymc.io/u/jessegrabowski)\
**Post date:** [November 27, 2024, 7:54am UTC](https://discourse.pymc.io/t/value-error-in-pytensor-function/16181/2 "2024-11-27T07:54:23Z")

</div>

All of your package versions are quite out of date. I think this stems from using python 3.8, which is no longer officially supported by pymc/pytensor. We try to follow the [numpy python version support timeline](https://numpy.org/neps/nep-0029-deprecation_policy.html), which dropped support for python versions less than 3.10 on April 05, 2024.

Current package versions are:

- pymc: 5.18.2
- pytensor: 2.26.3

I strongly recommend you upgrade (following the [official installation instructions](https://www.pymc.io/projects/docs/en/latest/installation.html)) and try again. I was able to run the example notebook without errors on the latest versions.