迁移学习 ValueError: Input 0 of layer

迁移学习 ValueError: Input 0 of layer global_average_pooling2d is incompatible with the layer: expected ndim=4, found ndim=2. Full shape received: [None, 1000]：

问题描述：

include_top要为False

    x2 = ResNet50(include_top=False, weights='imagenet', input_tensor=input_rgb)
    for layer in x2.layers:
        layer.trainable = False

    x2 = x2.output

    x2 = Flatten()(x2)
    x2 = Dense(512, activation='relu')(x2)
    x2 = Dropout(0.5)(x2)
    x2 = Dense(128, activation='relu')(x2)
    x2 = Dropout(0.5)(x2)
    x2 = Dense(64, activation='relu')(x2)
    x_2 = Model(inputs=input_rgb, outputs=x2)