发帖
雷达卡
这几天在看有关deep learning with r这本书，在看到CNN网络做图片识别那一章节，调试代码的时候发现，使用keras包中的fit_generator函数无法读取输入的图片生成器（train_generator），报错“UnidentifiedImageError:cannot identify image file"
查了很久也不知道如何解决这个问题，于是求问大佬们，是否有替代方案？
输入图片的格式是150×150，位深为3
以下是原代码：
train_datagen <- image_data_generator(rescale = 1/255)
validation_datagen <- image_data_generator(rescale = 1/255)
train_generator <- flow_images_from_directory(
train_dir, # target directory
train_datagen, # training data generator
target_size = c(150, 150), # resize all images
batch_size = 60,
class_mode = "binary" # because we use binary_crossentropy loss, we need binary labels
)
validation_generator <- flow_images_from_directory(
validation_dir,
validation_datagen,
target_size = c(150, 150),
batch_size = 60,
class_mode = "binary"
)
# output of thee generators is batches of 150x150 RGB images (shape(20, 150, 150, 3)) and binary labels(shape(20))
history <- model %>% fit_generator(
train_generator,
steps_per_epoch = 5,
epochs = 20,
validation_data = validation_generator,
validation_steps = 5
)
京公网安备 11010802022788号
