c.to_qiskit()失败

[UESTC-YLH]

你好！我在使用tensorcircuit进行mnist分类过程中，想将量子线路转化为qiskit线路，并且进行真机模拟。但是遇到以下错误：
File "/home/cpu_user_cpu/ylh_temporary/QUEST_IBMQ/mnist_2_qibm.py", line 67, in train_circuit
quantum_model.train(x_train, y_train)
File "/home/cpu_user_cpu/ylh_temporary/QUEST_IBMQ/model_file/model.py", line 163, in train
self.model.fit(x_train, y_train_one_hot, batch_size=self.batch_size, epochs=self.epochs, validation_split=0.2)
File "/home/cpu_user_cpu/condaenvstc/lib/python3.9/site-packages/keras/src/utils/traceback_utils.py", line 122, in error_handler
raise e.with_traceback(filtered_tb) from None
File "/home/cpu_user_cpu/condaenvstc/lib/python3.9/site-packages/tensorcircuit/keras.py", line 122, in call
result = tf.vectorized_map(
File "/home/cpu_user_cpu/condaenvstc/lib/python3.9/site-packages/tensorcircuit/keras.py", line 123, in
lambda vec: self.f(vec, *self.pqc_weights, **kwargs),
File "/home/cpu_user_cpu/ylh_temporary/QUEST_IBMQ/model_file/model.py", line 89, in _qml
qc = c.to_qiskit()
File "/home/cpu_user_cpu/condaenvstc/lib/python3.9/site-packages/tensorcircuit/abstractcircuit.py", line 770, in to_qiskit
return qir2qiskit(
File "/home/cpu_user_cpu/condaenvstc/lib/python3.9/site-packages/tensorcircuit/translation.py", line 295, in qir2qiskit
getattr(qiskit_circ, gate_name)(_get_float(parameters, "theta"), *index)
File "/home/cpu_user_cpu/condaenvstc/lib/python3.9/site-packages/tensorcircuit/translation.py", line 71, in _get_float
return np.real(backend.numpy(gates.array_to_tensor(parameters.get(key, default)))).item() # type: ignore
File "/home/cpu_user_cpu/condaenvstc/lib/python3.9/site-packages/tensorcircuit/backends/tensorflow_backend.py", line 475, in numpy
return a.numpy() # only valid in eager mode
AttributeError: Exception encountered when calling QuantumLayer.call().

Could not automatically infer the output shape / dtype of 'quantum_layer' (of type QuantumLayer). Either the QuantumLayer.call() method is incorrect, or you need to implement the QuantumLayer.compute_output_spec() / compute_output_shape() method. Error encountered:

'SymbolicTensor' object has no attribute 'numpy'

Arguments received by QuantumLayer.call():
• args=('<KerasTensor shape=(64, 10), dtype=float32, sparse=False, name=keras_tensor>',)
• kwargs={'mask': 'None'}

Process finished with exit code 1

我猜测问题可能出现在张量图转化失败，但是不知道应该怎么解决。在这里请求您的帮助！

[UESTC-YLH]

我不是很明白，您能帮我解释一下吗