keras自定义ImageDataGenerator

Hao Chen

无论如何,我们是否可以向ImageDataGenerator添加一些功能,以便ImageDataGenerator可以获取文件名列表以及每个小批量的随机样本图像?

我知道我可以自定义一个继承ImageDataGenerator类的类,但是我仍然不知道如何做到这一点。

这是我所做的:

for epoch in range(epochs):
    print("epoch is: %d, total epochs: %f" % ((epoch+1), int(epochs)))
    print("prepare training batch...")
    train_batch = makebatch(filelist=self.train_files, img_num=img_num, slice_times=slice_times)
    print("prepare validation batch..")
    val_batch = makebatch(filelist=self.val_files, img_num=int(math.ceil(img_num*0.2)), slice_times=slice_times)

    x_train = train_batch
    y_train = x_train

    x_val = val_batch
    y_val = x_val

    print("generate training data...")

    train_datagen.fit(x_train)
    train_generator = train_datagen.flow(
        x=x_train,
        y=y_train,
        batch_size=16)
    val_datagen.fit(x_val)
    val_generator = val_datagen.flow(
        x=x_val,
        y=y_val,
        batch_size=16)

    print("start training..")

    history = model.fit_generator(
            generator=train_generator,
            steps_per_epoch=None,
            epochs=1,
            verbose=1,
            validation_data=val_generator,
            validation_steps=None,
            callbacks=self.callbacks)

我真正想要获得的是,我可以为每个批次删除for循环和生成器随机样本图像。

有人可以帮忙吗?

talos1904

在这里,我会做的。

假设我有一个存储在变量X_train,X_validation中的所有图像的路径列表,并且标签存储为y_train和y_validation。

首先,我将定义一个序列生成器。(这是来自keras网站)

from skimage.io import imread
from skimage.transform import resize
import numpy as np

# Here, `x_set` is list of path to the images
# and `y_set` are the associated classes.    

class CIFAR10Sequence(Sequence):    
    def __init__(self, x_set, y_set, batch_size):
        self.x, self.y = x_set, y_set
        self.batch_size = batch_size

    def __len__(self):
        return int(np.ceil(len(self.x) / float(self.batch_size)))

    def __getitem__(self, idx):
        batch_x = self.x[idx * self.batch_size:(idx + 1) * self.batch_size]
        batch_y = self.y[idx * self.batch_size:(idx + 1) * self.batch_size]

        return np.array([
            resize(imread(file_name), (200, 200))
               for file_name in batch_x]), np.array(batch_y)

现在,我将用于训练和验证的生成器定义为

Xtrain_gen = detracSequence(X_train,y_train,batch_size=512) # you can choose your batch size.
Xvalidation_gen = detracSequence(X_validation,y_validation,batch_size=512)

现在,训练模型的最后一步

model.fit_generator(generator=Xtrain_gen, epochs=100, validation_data=Xvalidation_gen,use_multiprocessing=True)

这将为您避免for循环,并且这非常有效,因为CPU并行获取数据。

本文收集自互联网,转载请注明来源。

如有侵权,请联系 [email protected] 删除。

编辑于
0

我来说两句

0 条评论
登录 后参与评论

相关文章