TypeError:在RandomizedSearchCV上运行.fit（）时无法pickle _thread.RLock对象

Question

我正在尝试在RandomizedSearchCV上运行模型，以获得最佳的超参数

model = keras.models.Sequential([keras.layers.Dense(300, input_dim=7500, activation="relu"),
                                 keras.layers.Dense(300, activation = "relu", name = "Intermediate"),
                                 keras.layers.Dense(100, activation = "relu"),
                                 keras.layers.Dense(4, activation="softmax")])

keras_cls = keras.wrappers.scikit_learn.KerasClassifier(model)
    
    
param_distribs = {
    "n_hidden": [0, 1, 2, 3],
    "n_neurons": np.arange(1, 300),
    "learning_rate": 1e-3
}

rnd_search_cv = RandomizedSearchCV(keras_cls, param_distribs, n_iter=10, cv=3)
rnd_search_cv.fit(x_train, y_train, epochs=30,
                  validation_data=(x_test, y_test),
                  callbacks=[keras.callbacks.EarlyStopping(patience=10)])

在rmd_search_cv.fit我得到了错误TypeError: can't pickle _thread.RLock objects

应该采取什么措施来纠正错误