Как реализовать раннюю остановку в тензорном потоке

def train():
# Model
model = Model()

# Loss, Optimizer
global_step = tf.Variable(1, dtype=tf.int32, trainable=False, name='global_step')
loss_fn = model.loss()
optimizer = tf.train.AdamOptimizer(learning_rate=TrainConfig.LR).minimize(loss_fn, global_step=global_step)

# Summaries
summary_op = summaries(model, loss_fn)

with tf.Session(config=TrainConfig.session_conf) as sess:

    # Initialized, Load state
    sess.run(tf.global_variables_initializer())
    model.load_state(sess, TrainConfig.CKPT_PATH)

    writer = tf.summary.FileWriter(TrainConfig.GRAPH_PATH, sess.graph)

    # Input source
    data = Data(TrainConfig.DATA_PATH)

    loss = Diff()
    for step in xrange(global_step.eval(), TrainConfig.FINAL_STEP):

            mixed_wav, src1_wav, src2_wav, _ = data.next_wavs(TrainConfig.SECONDS, TrainConfig.NUM_WAVFILE, step)

            mixed_spec = to_spectrogram(mixed_wav)
            mixed_mag = get_magnitude(mixed_spec)

            src1_spec, src2_spec = to_spectrogram(src1_wav), to_spectrogram(src2_wav)
            src1_mag, src2_mag = get_magnitude(src1_spec), get_magnitude(src2_spec)

            src1_batch, _ = model.spec_to_batch(src1_mag)
            src2_batch, _ = model.spec_to_batch(src2_mag)
            mixed_batch, _ = model.spec_to_batch(mixed_mag)

            # Initializae our callback.
            #early_stopping_cb = EarlyStoppingCallback(val_acc_thresh=0.5)


            l, _, summary = sess.run([loss_fn, optimizer, summary_op],
                                     feed_dict={model.x_mixed: mixed_batch, model.y_src1: src1_batch,
                                                model.y_src2: src2_batch})

            loss.update(l)
            print('step-{}\td_loss={:2.2f}\tloss={}'.format(step, loss.diff * 100, loss.value))

            writer.add_summary(summary, global_step=step)

            # Save state
            if step % TrainConfig.CKPT_STEP == 0:
                tf.train.Saver().save(sess, TrainConfig.CKPT_PATH + '/checkpoint', global_step=step)

    writer.close()

У меня есть этот код нейронной сети, который отделяет музыку от голоса в WAV файле. как я могу ввести алгоритм ранней остановки для остановки секции поезда? Я вижу какой-то проект, который говорит о ValidationMonitor. Кто-нибудь может мне помочь?

ОТВЕТЫ

Ответ 1

ValidationMonitor отмечен как устаревший. это не рекомендуется. но вы все еще можете использовать его. вот пример того, как его создать:

    validation_monitor = monitors.ValidationMonitor(
        input_fn=functools.partial(input_fn, subset="evaluation"),
        eval_steps=128,
        every_n_steps=88,
        early_stopping_metric="accuracy",
        early_stopping_rounds = 1000
    )

и вы можете реализовать самостоятельно, здесь моя моя реализация:

          if (loss_value < self.best_loss):
            self.stopping_step = 0
            self.best_loss = loss_value
          else:
            self.stopping_step += 1
          if self.stopping_step >= FLAGS.early_stopping_step:
            self.should_stop = True
            print("Early stopping is trigger at step: {} loss:{}".format(global_step,loss_value))
            run_context.request_stop()

Ответ 2

Вот моя реализация ранней остановки, которую вы можете адаптировать:

Ранняя остановка может быть применена на определенных этапах тренировочного процесса, например, в конце каждой эпохи. В частности, в моем случае; Я отслеживаю потерю теста (проверки) в каждую эпоху, и после того, как потеря теста не улучшилась после 20 эпох (self.require_improvement= 20), обучение прерывается.

Вы можете установить максимальные эпохи на 10000 или 20000 или что угодно (self.max_epochs = 10000).

  self.require_improvement= 20
  self.max_epochs = 10000

Вот моя функция тренировки, где я использую раннюю остановку:

def train (self):

# training data
    train_input = self.Normalize(self.x_train)
    train_output = self.y_train.copy()            
#===============
    save_sess=self.sess # this used to compare the result of previous sess with actual one
# ===============
  #costs history :
    costs = []
    costs_inter=[]
# =================
  #for early stopping :
    best_cost=1000000 
    stop = False
    last_improvement=0
# ================
    n_samples = train_input.shape[0] # size of the training set
# ===============
   #train the mini_batches model using the early stopping criteria
    epoch = 0
    while epoch < self.max_epochs and stop == False:
        #train the model on the traning set by mini batches
        #suffle then split the training set to mini-batches of size self.batch_size
        seq =list(range(n_samples))
        random.shuffle(seq)
        mini_batches = [
            seq[k:k+self.batch_size]
            for k in range(0,n_samples, self.batch_size)
        ]

        avg_cost = 0. # The average cost of mini_batches
        step= 0

        for sample in mini_batches:

            batch_x = x_train.iloc[sample, :]
            batch_y =train_output.iloc[sample, :]
            batch_y = np.array(batch_y).flatten()

            feed_dict={self.X: batch_x,self.Y:batch_y, self.is_train:True}

            _, cost,acc=self.sess.run([self.train_step, self.loss_, self.accuracy_],feed_dict=feed_dict)
            avg_cost += cost *len(sample)/n_samples 
            print('epoch[{}] step [{}] train -- loss : {}, accuracy : {}'.format(epoch,step, avg_cost, acc))
            step += 100

        #cost history since the last best cost
        costs_inter.append(avg_cost)

        #early stopping based on the validation set/ max_steps_without_decrease of the loss value : require_improvement
        if avg_cost < best_cost:
            save_sess= self.sess # save session
            best_cost = avg_cost
            costs +=costs_inter # costs history of the validatio set
            last_improvement = 0
            costs_inter= []
        else:
            last_improvement +=1
        if last_improvement > self.require_improvement:
            print("No improvement found during the ( self.require_improvement) last iterations, stopping optimization.")
            # Break out from the loop.
            stop = True
            self.sess=save_sess # restore session with the best cost

        ## Run validation after every epoch : 
        print('---------------------------------------------------------')
        self.y_validation = np.array(self.y_validation).flatten()
        loss_valid, acc_valid = self.sess.run([self.loss_,self.accuracy_], 
                                              feed_dict={self.X: self.x_validation, self.Y: self.y_validation,self.is_train: True})
        print("Epoch: {0}, validation loss: {1:.2f}, validation accuracy: {2:.01%}".format(epoch + 1, loss_valid, acc_valid))
        print('---------------------------------------------------------')

        epoch +=1

Мы можем возобновить важный код здесь:

def train(self):
  ...
      #costs history :
        costs = []
        costs_inter=[]
      #for early stopping :
        best_cost=1000000 
        stop = False
        last_improvement=0
       #train the mini_batches model using the early stopping criteria
        epoch = 0
        while epoch < self.max_epochs and stop == False:
            ...
            for sample in mini_batches:
            ...                   
            #cost history since the last best cost
            costs_inter.append(avg_cost)

            #early stopping based on the validation set/ max_steps_without_decrease of the loss value : require_improvement
            if avg_cost < best_cost:
                save_sess= self.sess # save session
                best_cost = avg_cost
                costs +=costs_inter # costs history of the validatio set
                last_improvement = 0
                costs_inter= []
            else:
                last_improvement +=1
            if last_improvement > self.require_improvement:
                print("No improvement found during the ( self.require_improvement) last iterations, stopping optimization.")
                # Break out from the loop.
                stop = True
                self.sess=save_sess # restore session with the best cost
            ...
            epoch +=1

Надеюсь, это кому-нибудь поможет :).

Ответ 3

Начиная с версии r1.10, для API early_stopping.py доступны ранние r1.10 остановки в early_stopping.py (см. Github).

Например, tf.contrib.estimator.stop_if_no_decrease_hook (см. Документы)