Unknown image file format

TensorFlow General Discussion
,
1

Hello, I have a question. My image format is completely fine, but I keep getting an error saying “Unknown image file format. One of JPEG, PNG, GIF, BMP required.” Can someone tell me where I went wrong with my code? Here is my code:

import tensorflow as tf
from tensorflow import keras
from tensorflow.keras.optimizers import Adam #优化器
data_dir='Flug'
img_height=224
img_width=224


train_ds=tf.keras.utils.image_dataset_from_directory(
  data_dir,
  validation_split=0.2,
  subset="training",
  seed=123,
  image_size=(img_height, img_width),
  batch_size=10,
)
#验证集
val_ds = tf.keras.utils.image_dataset_from_directory(
  data_dir,
  validation_split=0.2,
  subset="validation",
  seed=123,
  image_size=(img_height, img_width),
  batch_size=10,
)

model = tf.keras.Sequential([
  tf.keras.layers.Rescaling(1./255), #数据在进入层序列时就已经通过这行代码归一化了
  tf.keras.layers.Conv2D(32, 3, activation='relu',input_shape=(224,224,3)),
  tf.keras.layers.MaxPooling2D(),
  tf.keras.layers.Flatten(),
  tf.keras.layers.Dense(64, activation='relu'),
  tf.keras.layers.Dense(1)
])
#model.build(input_shape=(None, img_height, img_width, 3))
#print(model.summary())

model.compile(loss='binary_crossentropy',optimizer=Adam(learning_rate=1e-4),metrics=['acc'])

model.fit(
  train_ds,
  validation_data=val_ds,
  epochs=10
)
3

@yunpeng_huo,

Welcome to the Tensorflow Forum!

When working with lots of real-world image data, corrupted images are a common occurrence. You can filter out badly-encoded images that do not feature the string “JFIF” in their header.

Please refer to the implementation part here.

Thank you!