有没有办法将MobileNet的输入维度从(224,224,3)修改为(150,150,3)

Question

有没有办法修改MobileNet的输入维度。每当我将其更改为所需的输入 (150,150,3) 时，它都会引发错误。

import tensorflow_hub as hub 
from tensorflow.keras import Sequential
# from tensorflow.keras import Activations
 classifier_url 
="https://hub.tensorflow.google.cn/google/tf2- 
 preview/mobilenet_v2/feature_vector/4"
baseModel = hub.KerasLayer(classifier_url, 
 input_shape=(150,150,3), output_shape=[1280], 
  name="Mobilenet")
 baseModel.trainable = False # freeze mobilenet 
  weights
 myModel = 
 Sequential(name="Mobilenet_tranferLearning")
 myModel.add(baseModel)
 myModel.add(Dropout(0.5))
  myModel.add(tf.keras.layers.Activation("relu"))
 myModel.add(Dense(102))
 myModel.add(tf.keras.layers.Activation("softmax"))
 myModel.summary()

ValueError：调用层“Mobilenet”（类型 KerasLayer）时遇到异常。

在用户代码中：

File "/usr/local/lib/python3.7/dist-packages/tensorflow_hub/keras_layer.py", line 237, in call  *
    result = smart_cond.smart_cond(training,

ValueError: Could not find matching concrete function to call loaded from the SavedModel. Got:
  Positional arguments (4 total):
    * Tensor("inputs:0", shape=(None, 150, 150, 3), dtype=float32)
    * False
    * False
    * 0.99
  Keyword arguments: {}

 Expected these arguments to match one of the following 4 option(s):

Option 1:
  Positional arguments (4 total):
    * TensorSpec(shape=(None, 224, 224, 3), dtype=tf.float32, name='inputs')
    * True
    * False
    * TensorSpec(shape=(), dtype=tf.float32, name='batch_norm_momentum')
  Keyword arguments: {}

Option 2:
  Positional arguments (4 total):
    * TensorSpec(shape=(None, 224, 224, 3), dtype=tf.float32, name='inputs')
    * True
    * True
    * TensorSpec(shape=(), dtype=tf.float32, name='batch_norm_momentum')
  Keyword arguments: {}

Option 3:
  Positional arguments (4 total):
    * TensorSpec(shape=(None, 224, 224, 3), dtype=tf.float32, name='inputs')
    * False
    * True
    * TensorSpec(shape=(), dtype=tf.float32, name='batch_norm_momentum')
  Keyword arguments: {}

Option 4:
  Positional arguments (4 total):
    * TensorSpec(shape=(None, 224, 224, 3), dtype=tf.float32, name='inputs')
    * False
    * False
    * TensorSpec(shape=(), dtype=tf.float32, name='batch_norm_momentum')
  Keyword arguments: {}

收到的调用参数： 输入= tf.Tensor（形状=（无，150，150，3），dtype = float32） • 训练=False

Answer 1

请查看此处，了解有关使用 MobilenetV2 的信息。
如果设置 include top=False，则不限于 224 X 224。另外，我建议您设置参数池 - 'Max'，这样 MobileNet 会输出一个向量，您可以直接将其输入到密集层中。下面是我用于数百个分类任务的代码。通常我使用 EfficientNetB3，但它也适用于 MobileNet。

import tensorflow as tf
from tensorflow import keras
from tensorflow.keras.layers import Dense, Dropout, BatchNormalization
from tensorflow.keras.optimizers import Adam, Adamax
from tensorflow.keras.metrics import categorical_crossentropy
from tensorflow.keras import regularizers
from tensorflow.keras.models import Model, load_model

def make_model(img_img_size, class_count,lr=.001, trainable=True):
    img_shape=(img_size[0], img_size[1], 3)    
    base_model=tf.keras.applications.MobileNet(include_top=False, 
                weights="imagenet",input_shape=img_shape, pooling='max') 
    base_model.trainable=trainable  
    x=base_model.output
    x=keras.layers.BatchNormalization(axis=-1, momentum=0.99, epsilon=0.001 )(x)
    x = Dense(256, kernel_regularizer = regularizers.l2(l = 0.016), 
              activity_regularizer = regularizers.l1(0.006),
              bias_regularizer = regularizers.l1(0.006) ,activation='relu')(x)
    x=Dropout(rate=.45, seed=123)(x)        
    output=Dense(class_count, activation='softmax')(x)
    model=Model(inputs=base_model.input, outputs=output)       
    model.compile(Adamax(learning_rate=lr), loss='categorical_crossentropy', metrics=['accuracy']) 
    return model, base_model # return the base_model so you can control its training state
img_size=(150, 150, 3)
class_count= 102 # looks like you have 102 classes
lr = .001
trainable=True # They always tell you to make the base  model not trainable. 
               # Train for a number of epochs than make the base model trainable.
               # I find it is better to make the base model trainable from the outset
               # It will converge faster and yield a lower validation loss
               # for the same number of epochs
   
model, base_model= make_model(img_img_size, class_count,lr=lr, trainable=trainable) 

注意 MobileNet 要求像素在 -1 到 +1 范围内缩放，因此请使用下面的函数来缩放输入图像。另外，我假设您的标签是一种热编码的标签，因此我将损失指定为“categorical_crossentropy”。如果这不是一个热门使用稀疏_分类_交叉熵。

def scaler(img):
    img=img/127.5-1
    return img