解决方案：

Question

所以我在keras上使用mnist示例，我正在尝试预测我自己的数字。我真的很难与尺寸大小匹配，因为我似乎无法找到一种方法来调整图像大小以使图像没有后面的行和列。我已尝试使用via numpy调整大小但是我在错误后得到错误...

代码

program Project1;
{$APPTYPE CONSOLE}

uses Types;

procedure A(i : integer; da : TDoubleDynArray); overload;
begin
end;

procedure A(s : string; da : TDoubleDynArray); overload;
begin
end;

begin
  A(1, [1.0]);
end.

错误

from __future__ import print_function
import keras
from keras.datasets import mnist
from keras.models import Sequential
from keras.layers import Dense, Dropout, Flatten
from keras.layers import Conv2D, MaxPooling2D
from keras import backend as K
import numpy as np
import cv2

batch_size = 20
num_classes = 10
epochs = 1
img_rows, img_cols = 28, 28

(x_train, y_train), (x_test, y_test) = mnist.load_data()

print("Processing image")   
im = cv2.imread('C:/Users/Luke/pic4.png', 0) #loading the image
print(im.shape) #28*28
im = cv2.resize(im,  (img_rows, img_cols)) 



list = [im]


batch = np.array([list for i in range(1)])   
print(batch.shape)#1*28*28
batch = batch.astype('float32')
batch /= 255

if K.image_data_format() == 'channels_first':
    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)
    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)
    input_shape = (1, img_rows, img_cols)
else:
    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)
    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)
    input_shape = (img_rows, img_cols, 1)

#print("x_train shape")     

x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
x_train /= 255
x_test /= 255

y_train = keras.utils.to_categorical(y_train, num_classes)
y_test = keras.utils.to_categorical(y_test, num_classes)

def base_model():
    model = Sequential()
    model.add(Conv2D(32, kernel_size=(3, 3),
                 activation='relu',
                 input_shape=input_shape))
    model.add(Conv2D(64, (3, 3), activation='relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))
    model.add(Dropout(0.25))
    model.add(Flatten())
    model.add(Dense(128, activation='relu'))
    model.add(Dropout(0.5))
    model.add(Dense(num_classes, activation='softmax'))

    model.compile(loss=keras.losses.categorical_crossentropy,
              optimizer=keras.optimizers.Adadelta(),
              metrics=['accuracy'])
return model

cnn_m = base_model()
cnn_m.summary()


print("Predicting image")
cnn_m.predict(batch)
print("Predicted image")

Answer 1

看起来你的数据格式错误了。您的数据以channels_first（即每张图像为1 x 28 x 28）传递，但Conv2D层需要channels_last（28 x 28 x 1）。

一个修复方法是将data_format=channels_first传递给Conv2D和MaxPooling图层。但是，如果您在CPU上运行，则可能不支持此功能。或者，更改此部分

if K.image_data_format() == 'channels_first':
    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)
    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)
    input_shape = (1, img_rows, img_cols)
else:
    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)
    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)
    input_shape = (img_rows, img_cols, 1)

始终执行else块（它重新整形为channels_last格式）。在这种情况下，请不要将data_format参数包含在Conv图层中（默认为channels_last）。

Answer 2

解决方案：

im = cv2.resize(im,  (img_rows, img_cols)) 
im.reshape((img_rows,img_cols))
print(im.shape) # (28,28)

batch = np.expand_dims(im,axis=0)
print(batch.shape) # (1, 28, 28)
batch = np.expand_dimes(batch,axis=3)
print(batch.shape) # (1, 28, 28,1)

... # build the model
model.predict(batch)

推理：

print(model.input_shape) # (None,28,28,1)

指任何批量大小（样本号），28 * 28形状和1个通道。在您的情况下，使用1作为样本号。

Answer 3

您只需添加

var getTest = (req, res, next) => {
    var categories = [];
    var promises = [];
    var names = ['Category 1', 'Category 2', 'Category 3', 'Category 4'];
    var resolveCount = 0;
    for (var i = 0; i < names.length; i++) {
        // Store the name in variable so that can be persistent and not
        // affected by the changing 'i' value
        const name = names[i]
        promises.push(new Promise((resolve, reject) => {
        
          // Your DB calls here. We shall use a simple timer to mimic the
          // effect
          setTimeout(() => {
            categories.push(name)
            resolveCount++;
            resolve();
          }, 1000)
        }));
    }
    
    
    Promise.all(promises).then(function() {
      console.log("This should run ONCE before AFTER promise resolved")
      console.log("resolveCount: " + resolveCount)
      console.log(categories);
      
      // Do your logic with the updated array
      // res.json(categories);
    });
    
    console.log("This will run immediately, before any promise resolve")
    console.log("resolveCount: " + resolveCount)
    console.log(categories)
}

getTest();

期望conv2d_1_input具有形状（28,28,1）但是具有形状的数组（1,28,28）

3 个答案:

解决方案：

推理：