Circuit not runnig by this backend instance.

[enzokgomes]

When attempting to run the trainer.fit method with the provided datasets and early stopping parameters, the following error is encountered:

CircuitNotRunError: Circuit corresponding to ResultHandle('1d68d2f5-60a4-48af-b771-51637cd791b3', 0, 3, 'null') has not been run by this backend instance.

This error occurs within the pytket library when trying to retrieve the result of a quantum circuit that has not been executed. The error is raised from the get_result method in the pytket backend.

Steps to Reproduce

1. Define train_dataset and val_dataset.
2. Initialize the trainer object.
3. Call the fit method on the trainer object with the following parameters:

trainer.fit(train_dataset, val_dataset,
            early_stopping_criterion='acc',
            early_stopping_interval=5,
            minimize_criterion=False)

Expected Behavior
The fit method should execute the training process without errors, utilizing the provided datasets and early stopping parameters.

Actual Behavior
The fit method raises a CircuitNotRunError, indicating that a quantum circuit corresponding to a specific ResultHandle has not been executed by the backend instance. This is followed by a KeyError when trying to access the result from the cache.

Error Traceback

---------------------------------------------------------------------------
CircuitNotRunError                        Traceback (most recent call last)
File c:\Users\enzok\AppData\Local\Programs\Python\Python311\Lib\site-packages\pytket\extensions\qiskit\backends\aer.py:344, in _AerBaseBackend.get_result(self, handle, **kwargs)
    343 try:
--> 344     return super().get_result(handle)
    345 except CircuitNotRunError:

File c:\Users\enzok\AppData\Local\Programs\Python\Python311\Lib\site-packages\pytket\backends\backend.py:347, in Backend.get_result(self, handle, **kwargs)
    346     return cast(BackendResult, self._cache[handle]["result"])
--> 347 raise CircuitNotRunError(handle)

CircuitNotRunError: Circuit corresponding to ResultHandle('1d68d2f5-60a4-48af-b771-51637cd791b3', 0, 3, 'null') has not been run by this backend instance.

During handling of the above exception, another exception occurred:

KeyError                                  Traceback (most recent call last)
Cell In[8], line 1
----> 1 trainer.fit(train_dataset, val_dataset,
      2             early_stopping_criterion='acc',
      3             early_stopping_interval=5,
      4             minimize_criterion=False)

File c:\Users\enzok\AppData\Local\Programs\Python\Python311\Lib\site-packages\lambeq\training\quantum_trainer.py:204, in QuantumTrainer.fit(self, train_dataset, val_dataset, log_interval, eval_interval, eval_mode, early_stopping_criterion, early_stopping_interval, minimize_criterion, full_timing_report)
    191 def fit(self,
    192         train_dataset: Dataset,
...
    356         "result"
    357     ] = backres
    359 return cast(BackendResult, self._cache[handle]["result"])

KeyError: ResultHandle('1d68d2f5-60a4-48af-b771-51637cd791b3', 1, 3, 'null')

Environment
Python Version: 3.11
pytket Version: 1.32.0
lambeq Version: 0.4.3
Operating System: Windows

The code is shared on: https://github.com/enzokgomes/NLP-quantum-training

[neiljdo]

Hi @enzokgomes,

Can you downgrade to pytket-qiskit~=0.50.0 and see if that works?