:return: Number of non changeable parameters

:return: number of changeable parameters

:param matrix: (nb_instruments, input_size, max_lenght, 2) :return:

:return:

:param loss_options: :param model_id: modelName;modelParam;nbSteps :param work_on: :param opt_param: :param

Use to find the output shapes of several convolutional layers :param input_shape: Tuple[int]: ⚠ batch dim is NOT in shape ⚠ :par

:param ags: :param kwags: :return:

:param epoch: :param logs: :return:

:param epoch: :param logs: :return:

:param epoch: :param logs: :return:

:param epoch: :param logs: :return:

:return:

:return:

:param key: :return:

:param key: :return:

:return:

:return:

Plays the models notes generated by the model :param time_step: :return:

:param label: :param names: list of strings :return:

:param filename: :param res: :return:

:param input: :return: All the input of the neural network

:param nb_steps: :param matrix: :param notes_list: :param verbose: :return:

Print a "working" summary :return: nothing

Print the weights :return:

:param mono: :param matrix: (nb_instruments, input_size, max_length, 2) :return: max and min note

create a new model witht the same options as the saved model and then load the weights (if with_weights==True) :param id: :pa

:param instrument_order: The order of the instruments to remplace :param song_number: The number of the song in the dataset

:param instrument_order: The order of the instruments to remplace :param song_number: The number of the song in the dataset

:param max_length: :param no_duration: :param verbose: :param save_images: :param noise: :return:

Generate Midi file from the seed and the trained model :param length: Length of th generation :param save_images: To save the

Used to reset the weights of a sublayers in a model Since Tensorflow 2.0 it works without this function :return:

:param array: (nb_instruments, batch=1, nb_steps, step_length, size, channels) :return: array: (nb_instruments, size, length, channe

:param name: The name of the instrument :return: All the possible instruments to train on which are similar

---------- (nb_instrument * 2 ->) [[a, b], [c, d, e]] has to become (e ->) [[d, c, b], [a, nb_instruments]] And the UpSampling is done o

:param take_all_steps_rhythm: :param max_reward: :param cost_value: :param y_true_a: activation, no loss (batch, nb_instruments, nb_

:param save_pil: :param save_plt: :param array: (a, b, 3) # RGB :param folder_path: :param name: :return:

:param notes_range: To add the piano on the left of the image :param array: (nb_instruments, size, length, channels) :param folder_path:

:param arrays: List[(a, b, 3)] :param file_name: :param titles: :param subtitles: :return:

:param notes_range: To add the image of the piano on the left :param subtitles: :param titles: :param arrays: List[(nb_instruments,

Save the best results and the parameters in a folder path :param best_accuracy: :param param_dict: :param folder_path: :return:

:param search_result: :param folder_path: :return:

All the generated arrays in one subplot to give an easier way to compare them :param subtitles: :param titles: :para

:param array: shape (nb_instruments, 128, nb_steps, 2) :return:

:param output_notes_list: the notes :param instruments : le list of the name of the instruments used :param path: The .mid file path

:param path: path were to save the model, if None it will be at self.saved_model_path :return:

:param search_result: :param folder_path: :return:

:param folder_path: :param search_result: :param dimensions: :return:

Save the list of the sorted scores and the correspondings parameters :param default_params_dict: :param folder_path: :param sorted_sc

:param mono: :param data: :param path: :return:

:param y_true_a: activation, no loss (batch, nb_instruments, nb_steps, step_size, input_size) :param y_pred_a: activation, no loss (batch, n

Add the scale and rhythm reward/cost :param mono: :param take_all_steps_rhythm: :param l_scale: :param l_rhythm: :param args:

To see the output of MySequence class can use : see_MySequence(*mySequence[i]) :param x: :param y: :return:

:param inputs: (nb_instruments, batch=2, nb_steps, step_size, input_size, 2) :param outputs: (nb_instruments, batch=2, step_size, input_size

:param x: (nb_instruments, batch, nb_steps, step_size, inputs_size, 2) :param yt: (nb_instruments, batch, step_size, input_size, 2) :par

:param sequence: :return:

set the default parameters of default_dict in d :param d: :param default_dict: :return:

set the default parameters of default_dict in d Work recursively if a key of default dict is a dictionary too :param d: :param defau

Used to set the weights of a sublayers in a model Since Tensorflow 2.0 it works without this function :param args:

:param args: :param kwargs: :return:

:param name: Name of the instrument :return: The music21 corresponding instrument class to instenciate

Create a list from the tupple '0:10:2' -> [0, 2, 4, 6, 8] :param string: :return:

:param path: The folder where to save :param title: Title of the summary :param file_name: Name of the file (without .txt) :param d:

:return: A dict with all the interesting attribut of the instance

:param x: :return: 10 ** (-x)

transform a list of shapes by replacing the value of nb_steps to x :param l: list of the shapes :param axis: axis of the time steps

:param step_length: :return:

:param matrix: (nb_instruments, 88, nb_steps, 2) :return: