崩溃后恢复训练 ACGAN，keras

Question

你如何继续训练 GAN？需要保存什么才能这样做？

# define the standalone discriminator model
def define_discriminator(in_shape=(64,64,1), n_classes=2):
    # weight initialization
    init = RandomNormal(stddev=0.02)
    # image input
    in_image = Input(shape=in_shape)
    # downsample to 32x32
    fe = Conv2D(32, (3,3), strides=(2,2), padding='same', kernel_initializer=init)(in_image)
    fe = LeakyReLU(alpha=0.2)(fe)
    fe = Dropout(0.5)(fe)
    # normal
    fe = Conv2D(64, (3,3), padding='same', kernel_initializer=init)(fe)
    fe = BatchNormalization()(fe)
    fe = LeakyReLU(alpha=0.2)(fe)
    fe = Dropout(0.5)(fe)
    # downsample to 16x16
    fe = Conv2D(128, (3,3), strides=(2,2), padding='same', kernel_initializer=init)(fe)
    fe = BatchNormalization()(fe)
    fe = LeakyReLU(alpha=0.2)(fe)
    fe = Dropout(0.5)(fe)
    # normal
    fe = Conv2D(256, (3,3), padding='same', kernel_initializer=init)(fe)
    fe = BatchNormalization()(fe)
    fe = LeakyReLU(alpha=0.2)(fe)
    fe = Dropout(0.5)(fe)
    # flatten feature maps
    fe = Flatten()(fe)
    # real/fake output
    out1 = Dense(1, activation='sigmoid')(fe)
    # class label output
    out2 = Dense(n_classes, activation='softmax')(fe)
    # define model
    model = Model(in_image, [out1, out2])
    # compile model
    opt = Adam(lr=0.0002, beta_1=0.5)
    model.compile(loss=['binary_crossentropy', 'sparse_categorical_crossentropy'], optimizer=opt)
    return model

# define the standalone generator model
def define_generator(latent_dim, n_classes=2):
    # weight initialization
    init = RandomNormal(stddev=0.02)
    # label input
    in_label = Input(shape=(1,))
    # embedding for categorical input
    li = Embedding(n_classes, 50)(in_label)
    # linear multiplication
    n_nodes = 16 * 16
    li = Dense(n_nodes, kernel_initializer=init)(li)
    # reshape to additional channel
    li = Reshape((16, 16, 1))(li)
    # image generator input
    in_lat = Input(shape=(latent_dim,))
    # foundation for 16x16 image
    n_nodes = 384 * 16 * 16
    gen = Dense(n_nodes, kernel_initializer=init)(in_lat)
    gen = Activation('relu')(gen)
    gen = Reshape((16, 16, 384))(gen)
    # merge image gen and label input
    merge = Concatenate()([gen, li])
    # upsample to 32x32
    gen = Conv2DTranspose(192, (5,5), strides=(2,2), padding='same', kernel_initializer=init)(merge)
    gen = BatchNormalization()(gen)
    gen = Activation('relu')(gen)
    # upsample to 64x64
    gen = Conv2DTranspose(1, (5,5), strides=(2,2), padding='same', kernel_initializer=init)(gen)
    out_layer = Activation('tanh')(gen)
    # define model
    model = Model([in_lat, in_label], out_layer)
    return model

# define the combined generator and discriminator model, for updating the generator
def define_gan(g_model, d_model):
    # make weights in the discriminator not trainable
    d_model.trainable = False
    # connect the outputs of the generator to the inputs of the discriminator
    gan_output = d_model(g_model.output)
    # define gan model as taking noise and label and outputting real/fake and label outputs
    model = Model(g_model.input, gan_output)
    # compile model
    opt = Adam(lr=0.0002, beta_1=0.5)
    model.compile(loss=['binary_crossentropy', 'sparse_categorical_crossentropy'], optimizer=opt)
    return model

我尝试过的是，当再次调用 train 函数时，我像这样以 5 个时期的间隔保存了生成器、GAN 和鉴别器

    g_model = load_model(g_model_save_path)
    print("GModel Loaded")
    d_model = load_model(d_model_save_path)
    print("dModel loaded")
    opt = Adam(lr=0.0002, beta_1=0.5)
    d_model.compile(loss=['binary_crossentropy', 'sparse_categorical_crossentropy'], optimizer=opt)
    gan_model = load_model(gan_model_save_path)
    print("GANMOdel loaded")
    opt = Adam(lr=0.0002, beta_1=0.5)
    gan_model.compile(loss=['binary_crossentropy', 'sparse_categorical_crossentropy'], optimizer=opt)

但是恢复后的效果并不好，即使是新恢复的模型大小也小了很多。