ValueError：层顺序的输入 0 与层不兼容：输入形状的预期轴 -1 具有值 3，但接收到的输入具有形状

Posted 2023-02-16

【中文标题】ValueError：层顺序的输入 0 与层不兼容：输入形状的预期轴 -1 具有值 3，但接收到的输入具有形状

【英文标题】：ValueError: Input 0 of layer sequential is incompatible with the layer: expected axis -1 of input shape to have value 3 but received input with shape

【发布时间】：2022-01-01 13:09:58

【问题描述】：

我是该主题的新手，遇到错误，谁能帮助我？

尝试使用手工数据集识别形状。

我无法解决此错误：

ValueError：层顺序的输入 0 与层不兼容：输入形状的预期轴 -1 具有值 3，但接收到形状为 [None, 100, 100, 1] 的输入

这是我的代码：

from keras.models import Sequential
from keras.layers import Dense, Flatten, Dropout
from keras.layers import Convolution2D, MaxPooling2D
from keras.preprocessing.image import ImageDataGenerator
from keras.preprocessing.image import load_img,img_to_array

model1 = Sequential()
model1.add(Convolution2D(32, (3,3), activation='sigmoid',input_shape=(100,100,3)))
model1.add(MaxPooling2D(pool_size=(2,2)))
model1.add(Convolution2D(32,(3,3),activation='relu'))
model1.add(MaxPooling2D(pool_size=(2,2)))
model1.add(Convolution2D(64,(3,3),activation='relu'))
model1.add(Convolution2D(64,(3,3),activation='relu'))
model1.add(MaxPooling2D(pool_size=(2,2)))

model1.add(Flatten())
model1.add(Dense(64,activation='relu'))
model1.add(Dropout(0.2))
model1.add(Dense(1,activation='sigmoid'))
model1.compile(loss='binary_crossentropy',
              optimizer='adam',
              metrics=['accuracy'])

model1.summary()


train_datagen = ImageDataGenerator(
        width_shift_range=0.2, height_shift_range=0.2,
        rotation_range=40,
        zoom_range=0.2,
        horizontal_flip=True,
        rescale=1/255,
        fill_mode='nearest')


validation_datagen = ImageDataGenerator(
        rescale=0.1,
        fill_mode='nearest')

# trainning
train_generator = train_datagen.flow_from_directory(
        'data/train',
        target_size=(100, 100),
        batch_size=16,
        class_mode='binary')
validation_generator = validation_datagen.flow_from_directory(
        'data/validation',
        target_size=(100, 100),
        batch_size=16,
        class_mode='binary')


h = model1.fit_generator(
        train_generator,
        epochs=10,
        validation_data=validation_generator)




img = load_img('img.jpeg', color_mode="grayscale", target_size=(100, 100))
img = img_to_array(img)
img = img.reshape((1, img.shape[0], img.shape[1], img.shape[2]))

prediction = model1.predict(img)




'The output:'

Traceback (most recent call last):
  /custom_train.py", line 68, in <module>
    prediction = model1.predict(img)
  \lib\site-packages\tensorflow\python\keras\engine\training.py", line 130, in _method_wrapper
    return method(self, *args, **kwargs)
  \lib\site-packages\tensorflow\python\keras\engine\training.py", line 1599, in predict
    tmp_batch_outputs = predict_function(iterator)
  \lib\site-packages\tensorflow\python\eager\def_function.py", line 780, in __call__
    result = self._call(*args, **kwds)
  \lib\site-packages\tensorflow\python\eager\def_function.py", line 823, in _call
    self._initialize(args, kwds, add_initializers_to=initializers)
  \lib\site-packages\tensorflow\python\eager\def_function.py", line 696, in _initialize
    self._stateful_fn._get_concrete_function_internal_garbage_collected(  # pylint: disable=protected-access
  \lib\site-packages\tensorflow\python\eager\function.py", line 2855, in _get_concrete_function_internal_garbage_collected
    graph_function, _, _ = self._maybe_define_function(args, kwargs)
  \lib\site-packages\tensorflow\python\eager\function.py", line 3213, in _maybe_define_function
    graph_function = self._create_graph_function(args, kwargs)
  \lib\site-packages\tensorflow\python\eager\function.py", line 3065, in _create_graph_function
    func_graph_module.func_graph_from_py_func(
  \lib\site-packages\tensorflow\python\framework\func_graph.py", line 986, in func_graph_from_py_func
    func_outputs = python_func(*func_args, **func_kwargs)
  \lib\site-packages\tensorflow\python\eager\def_function.py", line 600, in wrapped_fn
    return weak_wrapped_fn().__wrapped__(*args, **kwds)
  \lib\site-packages\tensorflow\python\framework\func_graph.py", line 973, in wrapper
    raise e.ag_error_metadata.to_exception(e)
ValueError: in user code:

    \lib\site-packages\tensorflow\python\keras\engine\training.py:1462 predict_function  *
        return step_function(self, iterator)
    \lib\site-packages\tensorflow\python\keras\engine\training.py:1452 step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    \lib\site-packages\tensorflow\python\distribute\distribute_lib.py:1211 run
        return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
    \lib\site-packages\tensorflow\python\distribute\distribute_lib.py:2585 call_for_each_replica
        return self._call_for_each_replica(fn, args, kwargs)
    \lib\site-packages\tensorflow\python\distribute\distribute_lib.py:2945 _call_for_each_replica
        return fn(*args, **kwargs)
    \lib\site-packages\tensorflow\python\keras\engine\training.py:1445 run_step  **
        outputs = model.predict_step(data)
    \lib\site-packages\tensorflow\python\keras\engine\training.py:1418 predict_step
        return self(x, training=False)
    \lib\site-packages\tensorflow\python\keras\engine\base_layer.py:975 __call__
        input_spec.assert_input_compatibility(self.input_spec, inputs,
    \lib\site-packages\tensorflow\python\keras\engine\input_spec.py:212 assert_input_compatibility
        raise ValueError(

    ValueError: Input 0 of layer sequential is incompatible with the layer: expected axis -1 of input shape to have value 3 but received input with shape [None, 100, 100, 1]

``

【参考方案1】：

模型的输入形状有100x100x3的图像。所以这意味着图像有3个通道（RGB图像）。

但是，在这一行中：

img = load_img('img.jpeg', color_mode="grayscale", target_size=(100, 100))

您以灰度格式加载了 img，这意味着图像形状将是 (100x100x1)，这意味着只有一个通道。

尝试将该行更改为：

img = load_img('img.jpeg', color_mode="rgb", target_size=(100, 100))

Keras documentation 表示：

color_mode：“灰度”、“rgb”、“rgba”之一。默认值：“RGB”。图像是否将转换为 1、3 或 4 通道。

【讨论】：

好的，非常感谢，问题解决了！