import gc import json import os import sys from datetime import datetime import numpy as np import tensorflow as tf from tensorflow import keras from tqdm.auto import tqdm from .particle import Particle gpus = tf.config.experimental.list_physical_devices("GPU") if gpus: try: tf.config.experimental.set_memory_growth(gpus[0], True) except RuntimeError as r: print(r) class Optimizer: """ particle swarm optimization PSO 실행을 위한 클래스 """ def __init__( self, model: keras.models, loss="mean_squared_error", n_particles: int = 10, c0=0.5, c1=1.5, w_min=0.5, w_max=1.5, negative_swarm: float = 0, mutation_swarm: float = 0, np_seed: int = None, tf_seed: int = None, random_state: tuple = None, particle_min: float = -5, particle_max: float = 5, ): """ particle swarm optimization Args: model (keras.models): 모델 구조 - keras.models.model_from_json 을 이용하여 생성 loss (str): 손실함수 - keras.losses 에서 제공하는 손실함수 사용 n_particles (int): 파티클 개수 c0 (float): local rate - 지역 최적값 관성 수치 c1 (float): global rate - 전역 최적값 관성 수치 w_min (float): 최소 관성 수치 w_max (float): 최대 관성 수치 negative_swarm (float): 최적해와 반대로 이동할 파티클 비율 - 0 ~ 1 사이의 값 mutation_swarm (float): 돌연변이가 일어날 확률 np_seed (int, optional): numpy seed. Defaults to None. tf_seed (int, optional): tensorflow seed. Defaults to None. particle_min (float, optional): 가중치 초기화 최소값. Defaults to -5. particle_max (float, optional): 가중치 초기화 최대값. Defaults to 5. """ if np_seed is not None: np.random.seed(np_seed) if tf_seed is not None: tf.random.set_seed(tf_seed) self.random_state = np.random.get_state() if random_state is not None: np.random.set_state(random_state) model.compile(loss=loss, optimizer="sgd", metrics=["accuracy"]) self.model = model # 모델 구조 self.loss = loss # 손실함수 self.n_particles = n_particles # 파티클 개수 self.particles = [None] * n_particles # 파티클 리스트 self.c0 = c0 # local rate - 지역 최적값 관성 수치 self.c1 = c1 # global rate - 전역 최적값 관성 수치 self.w_min = w_min # 최소 관성 수치 self.w_max = w_max # 최대 관성 수치 self.negative_swarm = negative_swarm # 최적해와 반대로 이동할 파티클 비율 - 0 ~ 1 사이의 값 self.mutation_swarm = mutation_swarm # 관성을 추가로 사용할 파티클 비율 - 0 ~ 1 사이의 값 self.particle_min = particle_min # 가중치 초기화 최소값 self.particle_max = particle_max self.g_best_score = [0, np.inf] # 최고 점수 - 시작은 0으로 초기화 self.g_best = None # 최고 점수를 받은 가중치 self.g_best_ = None # 최고 점수를 받은 가중치 - 값의 분산을 위한 변수 self.avg_score = 0 # 평균 점수 self.sigma = 1.0 self.save_path = None # 저장 위치 self.renewal = "acc" self.dispersion = False self.day = datetime.now().strftime("%Y%m%d-%H%M%S") self.empirical_balance = False negative_count = 0 self.train_summary_writer = [None] * self.n_particles try: print(f"start running time : {self.day}") for i in tqdm(range(self.n_particles), desc="Initializing Particles"): model_ = keras.models.model_from_json(model.to_json()) w_, sh_, len_ = self._encode(model_.get_weights()) w_ = np.random.uniform(particle_min, particle_max, len(w_)) model_.set_weights(self._decode(w_, sh_, len_)) model_.compile(loss=self.loss, optimizer="sgd", metrics=["accuracy"]) self.particles[i] = Particle( model_, loss, negative=True if i < negative_swarm * self.n_particles else False, mutation=mutation_swarm, ) if i < negative_swarm * self.n_particles: negative_count += 1 # del m, init_weights, w_, sh_, len_ gc.collect() tf.keras.backend.reset_uids() tf.keras.backend.clear_session() print(f"negative swarm : {negative_count} / {self.n_particles}") print(f"mutation swarm : {mutation_swarm * 100}%") gc.collect() tf.keras.backend.reset_uids() tf.keras.backend.clear_session() except KeyboardInterrupt: sys.exit("Ctrl + C : Stop Training") except MemoryError: sys.exit("Memory Error : Stop Training") except Exception as e: sys.exit(e) def __del__(self): del self.model del self.loss del self.n_particles del self.particles del self.c0 del self.c1 del self.w_min del self.w_max del self.negative_swarm del self.g_best_score del self.g_best del self.g_best_ del self.avg_score gc.collect() tf.keras.backend.reset_uids() tf.keras.backend.clear_session() def _encode(self, weights): """ 가중치를 1차원으로 풀어서 반환 Args: weights (list) : keras model의 가중치 Returns: (numpy array) : 가중치 - 1차원으로 풀어서 반환 (list) : 가중치의 원본 shape (list) : 가중치의 원본 shape의 길이 """ w_gpu = np.array([]) length = [] shape = [] for layer in weights: shape.append(layer.shape) w_tmp = layer.reshape(-1) length.append(len(w_tmp)) w_gpu = np.append(w_gpu, w_tmp) del weights return w_gpu, shape, length def _decode(self, weight, shape, length): """ _encode 로 인코딩된 가중치를 원본 shape으로 복원 파라미터는 encode의 리턴값을 그대로 사용을 권장 Args: weight (numpy array): 가중치 - 1차원으로 풀어서 반환 shape (list): 가중치의 원본 shape length (list): 가중치의 원본 shape의 길이 Returns: (list) : 가중치 원본 shape으로 복원 """ weights = [] start = 0 for i in range(len(shape)): end = start + length[i] w_tmp = weight[start:end] w_tmp = np.reshape(w_tmp, shape[i]) weights.append(w_tmp) start = end del weight, shape, length del start, end, w_tmp return weights def _f(self, x, y, weights): """ EBPSO의 목적함수 (예상) Args: x (list): 입력 데이터 y (list): 출력 데이터 weights (list): 가중치 Returns: (float): 목적 함수 값 """ self.model.set_weights(weights) score = self.model.evaluate(x, y, verbose=0) if self.renewal == "acc": score_ = score[1] else: score_ = score[0] if score_ > 0: return 1 / (1 + score_) else: return 1 + np.abs(score_) def fit( self, x, y, epochs: int = 100, log: int = 0, log_name: str = None, save_info: bool = False, save_path: str = "./result", renewal: str = "acc", empirical_balance: bool = False, dispersion: bool = False, check_point: int = None, ): """ # Args: x : numpy array, y : numpy array, epochs : int, log : int - 0 : log 기록 안함, 1 : log, 2 : tensorboard, save_info : bool - 종료시 학습 정보 저장 여부 default : False, save_path : str - ex) "./result", renewal : str ex) "acc" or "loss" or "both", empirical_balance : bool - True : EBPSO, False : PSO, dispersion : bool - True : g_best 의 값을 분산시켜 전역해를 찾음, False : g_best 의 값만 사용 check_point : int - 저장할 위치 - None : 저장 안함 """ self.save_path = save_path self.empirical_balance = empirical_balance self.dispersion = dispersion self.renewal = renewal particle_sum = 0 # x_j try: train_log_dir = "logs/fit/" + self.day if log == 2: assert log_name is not None, "log_name is None" train_log_dir = f"logs/{log_name}/{self.day}/train" for i in range(self.n_particles): self.train_summary_writer[i] = tf.summary.create_file_writer( train_log_dir + f"/{i}" ) elif check_point is not None or log == 1: if save_path is None: raise ValueError("save_path is None") else: self.save_path = save_path if not os.path.exists(f"{save_path}/{self.day}"): os.makedirs(f"{save_path}/{self.day}", exist_ok=True) except ValueError as e: sys.exit(e) except Exception as e: sys.exit(e) for i in tqdm(range(self.n_particles), desc="Initializing velocity"): p = self.particles[i] local_score = p.get_score(x, y, renewal=renewal) particle_sum += local_score[1] if renewal == "acc": if local_score[1] > self.g_best_score[0]: self.g_best_score[0] = local_score[1] self.g_best_score[1] = local_score[0] self.g_best = p.get_best_weights() self.g_best_ = p.get_best_weights() elif renewal == "loss": if local_score[0] < self.g_best_score[1]: self.g_best_score[1] = local_score[0] self.g_best_score[0] = local_score[1] self.g_best = p.get_best_weights() self.g_best_ = p.get_best_weights() elif renewal == "both": if local_score[1] > self.g_best_score[0]: self.g_best_score[0] = local_score[1] self.g_best_score[1] = local_score[0] self.g_best = p.get_best_weights() self.g_best_ = p.get_best_weights() if log == 1: with open( f"./{save_path}/{self.day}/{self.n_particles}_{epochs}_{self.c0}_{self.c1}_{self.w_min}_{renewal}.csv", "a", ) as f: f.write(f"{local_score[0]}, {local_score[1]}") if i != self.n_particles - 1: f.write(", ") else: f.write("\n") elif log == 2: with self.train_summary_writer[i].as_default(): tf.summary.scalar("loss", local_score[0], step=0) tf.summary.scalar("accuracy", local_score[1], step=0) del local_score # gc.collect() # tf.keras.backend.reset_uids() # tf.keras.backend.clear_session() print( f"initial g_best_score : {self.g_best_score[0] if self.renewal == 'acc' else self.g_best_score[1]}" ) try: epoch_sum = 0 epochs_pbar = tqdm( range(epochs), desc=f"best {self.g_best_score[0]:.4f}|{self.g_best_score[1]:.4f}", ascii=True, leave=True, position=0, ) for epoch in epochs_pbar: particle_avg = particle_sum / self.n_particles # x_j particle_sum = 0 max_score = 0 min_loss = np.inf # epoch_particle_sum = 0 part_pbar = tqdm( range(len(self.particles)), desc=f"acc : {max_score:.4f} loss : {min_loss:.4f}", ascii=True, leave=False, position=1, ) w = self.w_max - (self.w_max - self.w_min) * epoch / epochs for i in part_pbar: part_pbar.set_description( f"acc : {max_score:.4f} loss : {min_loss:.4f}" ) g_best = self.g_best if dispersion: ts = self.particle_min + np.random.rand() * ( self.particle_max - self.particle_min ) g_, g_sh, g_len = self._encode(self.g_best) decrement = (epochs - epoch + 1) / epochs g_ = (1 - decrement) * g_ + decrement * ts self.g_best_ = self._decode(g_, g_sh, g_len) g_best = self.g_best_ if empirical_balance: if np.random.rand() < np.exp(-(epoch) / epochs): w_p_ = self._f(x, y, self.particles[i].get_best_weights()) w_g_ = self._f(x, y, self.g_best) w_p = w_p_ / (w_p_ + w_g_) w_g = w_p_ / (w_p_ + w_g_) del w_p_ del w_g_ else: p_b = self.particles[i].get_best_score() g_a = self.avg_score l_b = p_b - g_a sigma_post = np.sqrt(np.power(l_b, 2)) sigma_pre = ( 1 / ( self.n_particles * np.linalg.norm( self.particle_max - self.particle_min ) ) * sigma_post ) p_ = np.exp(-1 * sigma_pre * sigma_post) # p_ = ( # 1 # / (self.n_particles * np.linalg.norm(self.particle_max - self.particle_min)) # * np.exp( # -np.power(l_b, 2) / (2 * np.power(self.sigma, 2)) # ) # ) # g_ = ( # 1 # / np.linalg.norm(self.c1 - self.c0) # * np.exp( # -np.power(l_b, 2) / (2 * np.power(self.sigma, 2)) # ) # ) # w_p = p_ / (p_ + g_) # w_g = g_ / (p_ + g_) w_p = p_ w_g = 1 - p_ del p_b del g_a del l_b del p_ score = self.particles[i].step_w( x, y, self.c0, self.c1, w, g_best, w_p, w_g, renewal=renewal, ) epoch_sum += np.power(score[1] - particle_avg, 2) else: score = self.particles[i].step( x, y, self.c0, self.c1, w, g_best, renewal=renewal ) if log == 2: with self.train_summary_writer[i].as_default(): tf.summary.scalar("loss", score[0], step=epoch + 1) tf.summary.scalar("accuracy", score[1], step=epoch + 1) if renewal == "acc": if score[1] >= max_score: max_score = score[1] min_loss = score[0] if score[1] >= self.g_best_score[0]: if score[1] > self.g_best_score[0]: self.g_best_score[0] = score[1] self.g_best = self.particles[i].get_best_weights() else: if score[0] < self.g_best_score[1]: self.g_best_score[1] = score[0] self.g_best = self.particles[i].get_best_weights() epochs_pbar.set_description( f"best {self.g_best_score[0]:.4f} | {self.g_best_score[1]:.4f}" ) elif renewal == "loss": if score[0] <= min_loss: min_loss = score[0] max_score = score[1] if score[0] <= self.g_best_score[1]: if score[0] < self.g_best_score[1]: self.g_best_score[1] = score[0] self.g_best = self.particles[i].get_best_weights() else: if score[1] > self.g_best_score[0]: self.g_best_score[0] = score[1] self.g_best = self.particles[i].get_best_weights() epochs_pbar.set_description( f"best {self.g_best_score[0]:.4f} | {self.g_best_score[1]:.4f}" ) elif renewal == "both": if score[0] <= min_loss: min_loss = score[0] if score[1] >= self.g_best_score[0]: self.g_best_score[0] = score[1] self.g_best = self.particles[i].get_best_weights() epochs_pbar.set_description( f"best {self.g_best_score[0]:.4f} | {self.g_best_score[1]:.4f}" ) if score[1] >= max_score: max_score = score[1] if score[0] <= self.g_best_score[1]: self.g_best_score[1] = score[0] self.g_best = self.particles[i].get_best_weights() epochs_pbar.set_description( f"best {self.g_best_score[0]:.4f} | {self.g_best_score[1]:.4f}" ) particle_sum += score[1] if log == 1: with open( f"./{save_path}/{self.day}/{self.n_particles}_{epochs}_{self.c0}_{self.c1}_{self.w_min}_{renewal}.csv", "a", ) as f: f.write(f"{score[0]}, {score[1]}") if i != self.n_particles - 1: f.write(", ") else: f.write("\n") part_pbar.refresh() if check_point is not None: if epoch % check_point == 0: os.makedirs(f"./{save_path}/{self.day}", exist_ok=True) self._check_point_save(f"./{save_path}/{self.day}/ckpt-{epoch}") gc.collect() tf.keras.backend.reset_uids() tf.keras.backend.clear_session() except KeyboardInterrupt: print("Ctrl + C : Stop Training") except MemoryError: print("Memory Error : Stop Training") except Exception as e: print(e) finally: self.model_save(save_path) print("model save") if save_info: self.save_info(save_path) print("save info") return self.g_best_score def get_best_model(self): """ 최고 점수를 받은 모델을 반환 Returns: (keras.models): 모델 """ model = keras.models.model_from_json(self.model.to_json()) model.set_weights(self.g_best) model.compile(loss=self.loss, optimizer="sgd", metrics=["accuracy"]) return model def get_best_score(self): """ 최고 점수를 반환 Returns: (float): 점수 """ return self.g_best_score def get_best_weights(self): """ 최고 점수를 받은 가중치를 반환 Returns: (float): 가중치 """ return self.g_best def save_info(self, path: str = "./result"): """ 학습 정보를 저장 Args: path (str, optional): 저장 위치. Defaults to "./result". """ json_save = { "name": f"{self.day}/{self.n_particles}_{self.c0}_{self.c1}_{self.w_min}.h5", "n_particles": self.n_particles, "score": self.g_best_score, "c0": self.c0, "c1": self.c1, "w_min": self.w_min, "w_max": self.w_max, "loss_method": self.loss, "empirical_balance": self.empirical_balance, "dispersion": self.dispersion, "negative_swarm": self.negative_swarm, "mutation_swarm": self.mutation_swarm, "random_state_0": self.random_state[0], "random_state_1": self.random_state[1].tolist(), "random_state_2": self.random_state[2], "random_state_3": self.random_state[3], "random_state_4": self.random_state[4], "renewal": self.renewal, } with open( f"./{path}/{self.day}/{self.loss}_{self.g_best_score}.json", "a", ) as f: json.dump(json_save, f, indent=4) def _check_point_save(self, save_path: str = f"./result/check_point"): """ 중간 저장 Args: save_path (str, optional): checkpoint 저장 위치 및 이름. Defaults to f"./result/check_point". """ model = self.get_best_model() model.save_weights(save_path) def model_save(self, save_path: str = "./result"): """ 최고 점수를 받은 모델 저장 Args: save_path (str, optional): 모델의 저장 위치. Defaults to "./result". Returns: (keras.models): 모델 """ model = self.get_best_model() model.save( f"./{save_path}/{self.day}/{self.n_particles}_{self.c0}_{self.c1}_{self.w_min}.h5" ) return model