How to use keras model.reset_state in function?

Question

Hi I have a question the way to reset model in for loop.

To do a grid search to find the optimal model for the dataset, I made for loop with parameters.

Here's the structure of the model.

list_loss = []

# Define model
def Model():
    model = Sequential()
    model.add(Dense(~~~~))
    model.add(Dense(1, activation = 'sigmoid')

    model.compile(optimizer = 'adam', loss = ~~)
    return model


model = Model()

for epoch in epochs_:
    for batch in batchs_:
        # train model
        hist = model.fit(X_train, y_train)


        # save the loss

        list_loss.append(hist.history['loss'][-1])
        print('layers: {}, batch size: {}, epoch: {},  and loss: {}'.format(a, batch_size_, epochs_, hist.history['loss'][-1])

And here's the result. it looks like the previous loss give impact to the next one. How should I reset on each epoch?

Thank you in advance.