thekangchen / soundscape-emotion-recognition-models Goto Github PK

I am interested in running the code you have provided, but I have encountered some errors while attempting to set up the necessary environment. I believe that having a complete environment setup file (e.g., a requirements.txt or environment.yml) would greatly assist me in resolving these issues.

Could you please provide me with such a file, if available? This would include all the dependencies and versions of the packages required to run your code smoothly.

Thank you very much !

AttributeError Traceback (most recent call last)
Input In [24], in <cell line: 19>()
42 #VALENCE
43 model_CNN_valence = KerasRegressor(build_fn = build_cnn, verbose=True, epochs=53, batch_size=32)
---> 45 history_valence = model_CNN_valence.fit(normalized_pathes_train_valence, y_patches_train_valence)
47 CNN_y_predicted_valence = model_CNN_valence.predict(normalized_pathes_test_valence)
49 CNN_y_predicted_valence = np.array(average_output_of_patches(CNN_y_predicted_valence))

File ~/miniconda3/lib/python3.8/site-packages/keras/wrappers/scikit_learn.py:167, in BaseWrapper.fit(self, x, y, **kwargs)
163 else:
164 self.model = self.build_fn(**self.filter_sk_params(self.build_fn))
166 if (
--> 167 losses.is_categorical_crossentropy(self.model.loss)
168 and len(y.shape) != 2
169 ):
170 y = to_categorical(y)
172 fit_args = copy.deepcopy(self.filter_sk_params(Sequential.fit))

AttributeError: module 'keras.losses' has no attribute 'is_categorical_crossentropy'