KeyError: 'text_block_1/conv_block_1/filters_0_1'

[RaedMatti]

I'm doing a text classification with 5 trials, in the 3rd I'm getting this error

KeyError                                  Traceback (most recent call last)
Cell In [58], line 1
----> 1 clf.fit(X_train, y_train,epochs=2, callbacks=None, validation_split=0.2)

File /usr/local/lib/python3.8/site-packages/autokeras/tasks/text.py:160, in TextClassifier.fit(self, x, y, epochs, callbacks, validation_split, validation_data, **kwargs)
    103 def fit(
    104     self,
    105     x=None,
   (...)
    111     **kwargs
    112 ):
    113     """Search for the best model and hyperparameters for the AutoModel.
    114 
    115     It will search for the best model based on the performances on
   (...)
    158             validation loss values and validation metrics values (if applicable).
    159     """
--> 160     history = super().fit(
    161         x=x,
    162         y=y,
    163         epochs=epochs,
    164         callbacks=callbacks,
    165         validation_split=validation_split,
    166         validation_data=validation_data,
    167         **kwargs
    168     )
    169     return history

File /usr/local/lib/python3.8/site-packages/autokeras/auto_model.py:292, in AutoModel.fit(self, x, y, batch_size, epochs, callbacks, validation_split, validation_data, verbose, **kwargs)
    287 if validation_data is None and validation_split:
    288     dataset, validation_data = data_utils.split_dataset(
    289         dataset, validation_split
    290     )
--> 292 history = self.tuner.search(
    293     x=dataset,
    294     epochs=epochs,
    295     callbacks=callbacks,
    296     validation_data=validation_data,
    297     validation_split=validation_split,
    298     verbose=verbose,
    299     **kwargs
    300 )
    302 return history

File /usr/local/lib/python3.8/site-packages/autokeras/engine/tuner.py:193, in AutoTuner.search(self, epochs, callbacks, validation_split, verbose, **fit_kwargs)
    191 self._try_build(hp)
    192 self.oracle.update_space(hp)
--> 193 super().search(
    194     epochs=epochs, callbacks=new_callbacks, verbose=verbose, **fit_kwargs
    195 )
    197 # Train the best model use validation data.
    198 # Train the best model with enough number of epochs.
    199 if validation_split > 0 or early_stopping_inserted:

File /usr/local/lib/python3.8/site-packages/keras_tuner/engine/base_tuner.py:173, in BaseTuner.search(self, *fit_args, **fit_kwargs)
    171 while True:
    172     self.pre_create_trial()
--> 173     trial = self.oracle.create_trial(self.tuner_id)
    174     if trial.status == trial_module.TrialStatus.STOPPED:
    175         # Oracle triggered exit.
    176         tf.get_logger().info("Oracle triggered exit")

File /usr/local/lib/python3.8/site-packages/keras_tuner/engine/oracle.py:194, in Oracle.create_trial(self, tuner_id)
    192     values = None
    193 else:
--> 194     response = self.populate_space(trial_id)
    195     status = response["status"]
    196     values = response["values"] if "values" in response else None

File /usr/local/lib/python3.8/site-packages/autokeras/tuners/greedy.py:153, in GreedyOracle.populate_space(self, trial_id)
    151 for _ in range(self._max_collisions):
    152     hp_names = self._select_hps()
--> 153     values = self._generate_hp_values(hp_names)
    154     # Reached max collisions.
    155     if values is None:

File /usr/local/lib/python3.8/site-packages/autokeras/tuners/greedy.py:189, in GreedyOracle._generate_hp_values(self, hp_names)
    186 if hps.is_active(hp):
    187     # if was active and not selected, do nothing.
    188     if best_hps.is_active(hp.name) and hp.name not in hp_names:
--> 189         hps.values[hp.name] = best_hps.values[hp.name]
    190         continue
    191     # if was not active or selected, sample.

KeyError: 'text_block_1/conv_block_1/filters_0_1'

im using

clf = ak.TextClassifier(
    overwrite=True, max_trials=5)  

and

clf.fit(X_train, y_train,epochs=2, callbacks=None, validation_split=0.2)

[YASIRAQ]

same problem . is there any solution ?

[haifeng-jin]

Let's use #1812 for the discussion