From fe33b892bd3197cc1e257889b4a19c267d51516b Mon Sep 17 00:00:00 2001 From: prise6 Date: Sun, 10 Mar 2019 19:36:42 +0100 Subject: [PATCH] adam optimizer --- iss/models/SimpleAutoEncoder.py | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/iss/models/SimpleAutoEncoder.py b/iss/models/SimpleAutoEncoder.py index 06139be..3309084 100644 --- a/iss/models/SimpleAutoEncoder.py +++ b/iss/models/SimpleAutoEncoder.py @@ -2,7 +2,7 @@ from iss.models.AbstractModel import AbstractModel from keras.layers import Input, Dense, Conv2D, MaxPooling2D, UpSampling2D, Reshape, Flatten -from keras.optimizers import Adadelta +from keras.optimizers import Adadelta, Adam from keras.models import Model import numpy as np @@ -26,11 +26,11 @@ class SimpleAutoEncoder(AbstractModel): picture = Input(shape = input_shape) x = Flatten()(picture) - layer_1 = Dense(2000, activation = 'relu', name = 'enc_1')(x) + layer_1 = Dense(1000, activation = 'relu', name = 'enc_1')(x) layer_2 = Dense(100, activation = 'relu', name = 'enc_2')(layer_1) - layer_3 = Dense(30, activation = 'relu', name = 'enc_3')(layer_2) + layer_3 = Dense(50, activation = 'relu', name = 'enc_3')(layer_2) layer_4 = Dense(100, activation = 'relu', name = 'dec_1')(layer_3) - layer_5 = Dense(2000, activation = 'relu', name = 'dec_2')(layer_4) + layer_5 = Dense(1000, activation = 'relu', name = 'dec_2')(layer_4) # encoded network # x = Conv2D(1, (3, 3), activation = 'relu', padding = 'same', name = 'enc_conv_1')(picture) @@ -45,6 +45,7 @@ class SimpleAutoEncoder(AbstractModel): self.model = Model(picture, decoded) - optimizer = Adadelta(lr = self.lr, rho = 0.95, epsilon = None, decay = 0.0) + # optimizer = Adadelta(lr = self.lr, rho = 0.95, epsilon = None, decay = 0.0) + optimizer = Adam(lr = 0.001, beta_1=0.9, beta_2=0.999, epsilon=None, decay=0.0, amsgrad=False) self.model.compile(optimizer = optimizer, loss = 'binary_crossentropy')