add original python code

+# coding: utf-8
+try:
+    import urllib.request
+except ImportError:
+    raise ImportError('You should use Python 3.x')
+import os.path
+import gzip
+import pickle
+import os
+import numpy as np
+
+
+key_file = {
+    'train':'cifar10-train.gz',
+    'test':'cifar10-test.gz'
+}
+
+dataset_dir = os.path.dirname(os.path.abspath('/Users/HyeonJun/Desktop/simple_convnet/dataset'))
+save_file = dataset_dir + "/cifar10.pkl"
+
+train_num = 50000
+test_num = 10000
+img_dim = (3, 32, 32)
+img_size = 3072
+
+def _load_label(file_name):
+    file_path = dataset_dir + "/" + file_name
+
+    print("Converting " + file_name + " to NumPy Array ...")
+    with gzip.open(file_path, 'rb') as f:
+            labels = np.frombuffer(f.read(), np.uint8, offset=0)
+    labels = labels.reshape(-1, img_size+1)
+    labels = labels.T
+    print("Done")
+
+    return labels[0]
+
+def _load_img(file_name):
+    file_path = dataset_dir + "/" + file_name
+
+    print("Converting " + file_name + " to NumPy Array ...")
+    with gzip.open(file_path, 'rb') as f:
+            data = np.frombuffer(f.read(), np.uint8, offset=0)
+    data = data.reshape(-1, img_size+1)
+    data = np.delete(data, 0, 1)
+    print("Done")
+
+    return data
+
+def _convert_numpy():
+    dataset = {}
+    dataset['train_img'] =  _load_img(key_file['train'])
+    dataset['train_label'] = _load_label(key_file['train'])
+    dataset['test_img'] = _load_img(key_file['test'])
+    dataset['test_label'] = _load_label(key_file['test'])
+
+    return dataset
+
+def init_cifar10():
+    dataset = _convert_numpy()
+    print("Creating pickle file ...")
+    with open(save_file, 'wb') as f:
+        pickle.dump(dataset, f, -1)
+    print("Done!")
+
+def _change_one_hot_label(X):
+    T = np.zeros((X.size, 10))
+    for idx, row in enumerate(T):
+        row[X[idx]] = 1
+
+    return T
+
+def load_cifar10(normalize=True, flatten=True, one_hot_label=False):
+    """CIFAR-10データセットの読み込み
+
+    Parameters
+    ----------
+    normalize : 画像のピクセル値を0.0~1.0に正規化する
+    one_hot_label :
+        one_hot_labelがTrueの場合、ラベルはone-hot配列として返す
+        one-hot配列とは、たとえば[0,0,1,0,0,0,0,0,0,0]のような配列
+    flatten : 画像を一次元配列に平にするかどうか
+
+    Returns
+    -------
+    (訓練画像, 訓練ラベル), (テスト画像, テストラベル)
+    """
+    if not os.path.exists(save_file):
+        init_cifar10()
+
+    with open(save_file, 'rb') as f:
+        dataset = pickle.load(f)
+
+    if normalize:
+        for key in ('train_img', 'test_img'):
+            dataset[key] = dataset[key].astype(np.float32)
+            dataset[key] /= 255.0
+
+    if one_hot_label:
+        dataset['train_label'] = _change_one_hot_label(dataset['train_label'])
+        dataset['test_label'] = _change_one_hot_label(dataset['test_label'])
+
+    if not flatten:
+         for key in ('train_img', 'test_img'):
+            dataset[key] = dataset[key].reshape(-1, 3, 32, 32)
+
+    return (dataset['train_img'], dataset['train_label']), (dataset['test_img'], dataset['test_label'])
+
+
+if __name__ == '__main__':
+    init_cifar10()

+# coding: utf-8
+#import cupy as cp
+import numpy as cp
+import numpy as np
+
+
+def identity_function(x):
+    return x
+
+
+def step_function(x):
+    return np.array(x > 0, dtype=np.int)
+
+
+def sigmoid(x):
+    return 1 / (1 + np.exp(-x))    
+
+
+def sigmoid_grad(x):
+    return (1.0 - sigmoid(x)) * sigmoid(x)
+    
+
+def relu(x):
+    return np.maximum(0, x)
+
+
+def relu_grad(x):
+    grad = np.zeros(x)
+    grad[x>=0] = 1
+    return grad
+    
+
+def softmax(x):
+    if x.ndim == 2:
+        x = x.T
+        x = x - cp.max(x, axis=0)
+        y = cp.exp(x, dtype=np.float32) / cp.sum(cp.exp(x, dtype=np.float32), axis=0, dtype=np.float32)
+        return y.T 
+
+    x = x - cp.max(x) # オーバーフロー対策
+    return cp.exp(x) / cp.sum(cp.exp(x))
+
+
+def mean_squared_error(y, t):
+    return 0.5 * np.sum((y-t)**2)
+
+
+def cross_entropy_error(y, t):
+    if y.ndim == 1:
+        t = t.reshape(1, t.size)
+        y = y.reshape(1, y.size)
+        
+    # 教師データがone-hot-vectorの場合、正解ラベルのインデックスに変換
+    if t.size == y.size:
+        t = t.argmax(axis=1)
+             
+    batch_size = y.shape[0]
+    return -cp.sum(cp.log(y[cp.arange(batch_size), t])) / batch_size
+
+
+def softmax_loss(X, t):
+    y = softmax(X)
+    return cross_entropy_error(y, t)

+# coding: utf-8
+import numpy as np
+
+def _numerical_gradient_1d(f, x):
+    h = 1e-4 # 0.0001
+    grad = np.zeros_like(x)
+    
+    for idx in range(x.size):
+        tmp_val = x[idx]
+        x[idx] = float(tmp_val) + h
+        fxh1 = f(x) # f(x+h)
+        
+        x[idx] = tmp_val - h 
+        fxh2 = f(x) # f(x-h)
+        grad[idx] = (fxh1 - fxh2) / (2*h)
+        
+        x[idx] = tmp_val 
+        
+    return grad
+
+
+def numerical_gradient_2d(f, X):
+    if X.ndim == 1:
+        return _numerical_gradient_1d(f, X)
+    else:
+        grad = np.zeros_like(X)
+        
+        for idx, x in enumerate(X):
+            grad[idx] = _numerical_gradient_1d(f, x)
+        
+        return grad
+
+
+def numerical_gradient(f, x):
+    h = 1e-4 # 0.0001
+    grad = np.zeros_like(x)
+    
+    it = np.nditer(x, flags=['multi_index'], op_flags=['readwrite'])
+    while not it.finished:
+        idx = it.multi_index
+        tmp_val = x[idx]
+        x[idx] = float(tmp_val) + h
+        fxh1 = f(x) # f(x+h)
+        
+        x[idx] = tmp_val - h 
+        fxh2 = f(x) # f(x-h)
+        grad[idx] = (fxh1 - fxh2) / (2*h)
+        
+        x[idx] = tmp_val # 値を元に戻す
+        it.iternext()   
+        
+    return grad
\ No newline at end of file

+
+import numpy as cp
+import numpy as np
+
+class SGD:
+
+
+    def __init__(self, lr=0.01):
+        self.lr = lr
+        
+    def update(self, params, grads):
+        for key in params.keys():
+            params[key] -= self.lr * grads[key] 
+
+
+class Momentum:
+
+
+    def __init__(self, lr=0.01, momentum=0.9):
+        self.lr = lr
+        self.momentum = momentum
+        self.v = None
+        
+    def update(self, params, grads):
+        if self.v is None:
+            self.v = {}
+            for key, val in params.items():                                
+                self.v[key] = np.zeros_like(val)
+                
+        for key in params.keys():
+            self.v[key] = self.momentum*self.v[key] - self.lr*grads[key] 
+            params[key] += self.v[key]
+
+
+class Nesterov:
+
+
+    def __init__(self, lr=0.01, momentum=0.9):
+        self.lr = lr
+        self.momentum = momentum
+        self.v = None
+        
+    def update(self, params, grads):
+        if self.v is None:
+            self.v = {}
+            for key, val in params.items():
+                self.v[key] = np.zeros_like(val)
+            
+        for key in params.keys():
+            self.v[key] *= self.momentum
+            self.v[key] -= self.lr * grads[key]
+            params[key] += self.momentum * self.momentum * self.v[key]
+            params[key] -= (1 + self.momentum) * self.lr * grads[key]
+
+
+class AdaGrad:
+
+
+    def __init__(self, lr=0.01):
+        self.lr = lr
+        self.h = None
+        
+    def update(self, params, grads):
+        if self.h is None:
+            self.h = {}
+            for key, val in params.items():
+                self.h[key] = np.zeros_like(val)
+            
+        for key in params.keys():
+            self.h[key] += grads[key] * grads[key]
+            params[key] -= self.lr * grads[key] / (np.sqrt(self.h[key]) + 1e-7)
+
+
+class RMSprop:
+
+
+    def __init__(self, lr=0.01, decay_rate = 0.99):
+        self.lr = lr
+        self.decay_rate = decay_rate
+        self.h = None
+        
+    def update(self, params, grads):
+        if self.h is None:
+            self.h = {}
+            for key, val in params.items():
+                self.h[key] = np.zeros_like(val)
+            
+        for key in params.keys():
+            self.h[key] *= self.decay_rate
+            self.h[key] += (1 - self.decay_rate) * grads[key] * grads[key]
+            params[key] -= self.lr * grads[key] / (np.sqrt(self.h[key]) + 1e-7)
+
+
+class Adam:
+
+
+    def __init__(self, lr=0.001, beta1=0.9, beta2=0.999):
+        self.lr = lr
+        self.beta1 = beta1
+        self.beta2 = beta2
+        self.iter = 0
+        self.m = None
+        self.v = None
+        
+    def update(self, params, grads):
+        if self.m is None:
+            self.m, self.v = {}, {}
+            for key, val in params.items():
+                self.m[key] = cp.zeros_like(val, dtype=np.float32)
+                self.v[key] = cp.zeros_like(val, dtype=np.float32)
+        
+        self.iter += 1
+        lr_t  = self.lr * cp.sqrt(1.0 - self.beta2**self.iter, dtype=np.float32) / (1.0 - self.beta1**self.iter)
+        
+        for key in params.keys():
+           
+            self.m[key] += (1 - self.beta1) * (grads[key] - self.m[key])
+            self.v[key] += (1 - self.beta2) * (grads[key]**2 - self.v[key])
+            
+            params[key] -= lr_t * self.m[key] / (cp.sqrt(self.v[key], dtype=np.float32) + 1e-7)
+            
+           
+class EarlyStopping():
+    def __init__(self, patience=0, verbose=0):
+        self.step = 0
+        self.acc = 0.0
+        self.patience = patience
+        self.verbose = verbose
+
+    def validate(self, acc):
+        if self.acc > acc:
+            self.step += 1
+            if self.step > self.patience:
+                if self.verbose:
+                    print('early stopping')
+                return True
+        else:
+            self.step = 0
+            self.acc = acc
+        return False

+import pickle as pk
+import numpy as np
+
+with open('/Users/HyeonJun/Desktop/simple_convnet/params.pkl', 'rb') as file:
+	dict = pk.load(file)
+	for key in dict.keys():
+		print(key, " : ", dict[key])

+import sys, os
+sys.path.append(os.pardir)
+import pickle
+import numpy as cp
+import numpy as np
+from collections import OrderedDict
+from layers import *
+from gradient import numerical_gradient
+
+
+class SimpleConvNet:
+    def __init__(self, input_dim=(3, 32, 32),
+                 conv_param={'filter_num':(32, 32, 64), 'filter_size':3, 'pad':1, 'stride':1},
+                 hidden_size=512, output_size=10, weight_init_std=0.01):
+        filter_num = conv_param['filter_num']
+        filter_size = conv_param['filter_size']
+        filter_pad = conv_param['pad']
+        filter_stride = conv_param['stride']
+        input_size = input_dim[1]
+        conv_output_size = (input_size - filter_size + 2*filter_pad) / filter_stride + 1
+        conv_data_size    = int(filter_num[0] *  conv_output_size    *  conv_output_size   )
+        pool1_output_size = int(filter_num[1] * (conv_output_size/2) * (conv_output_size/2))
+        pool2_output_size = int(filter_num[2] * (conv_output_size/4) * (conv_output_size/4))
+        pool3_output_size = int(filter_num[2] * (conv_output_size/8) * (conv_output_size/8))
+
+        self.params = {}
+        self.params['W1'] = cp.array( weight_init_std * \
+                            cp.random.randn(filter_num[0], input_dim[0], filter_size, filter_size), dtype=np.float32)
+
+        self.params['W2'] = cp.array( weight_init_std * \
+                            cp.random.randn(filter_num[1], filter_num[0], 1, 1), dtype=np.float32)
+
+        self.params['W3'] = cp.array( weight_init_std * \
+                            cp.random.randn(filter_num[1], 1, filter_size, filter_size), dtype=np.float32)
+
+        self.params['W4'] = cp.array( weight_init_std * \
+                            cp.random.randn(filter_num[2], filter_num[1], 1, 1), dtype=np.float32)
+
+        self.params['W5'] = cp.array( weight_init_std * \
+                            cp.random.randn(filter_num[2], 1, filter_size, filter_size), dtype=np.float32)
+
+        self.params['W6'] = cp.array( weight_init_std * \
+                            cp.random.randn(pool3_output_size, hidden_size), dtype=np.float32)
+
+        self.params['W7'] = cp.array( weight_init_std * \
+                            cp.random.randn(hidden_size, output_size), dtype=np.float32)
+
+        self.layers = OrderedDict()
+        self.layers['Conv1'] = Convolution(self.params['W1'],
+                                           conv_param['stride'], conv_param['pad'])
+        self.layers['LightNorm1'] = LightNormalization()
+        self.layers['Relu1'] = Relu()
+        self.layers['Pool1'] = Pooling(pool_h=2, pool_w=2, stride=2)
+
+        self.layers['Conv2'] = Convolution(self.params['W2'],
+                                           1, 0)
+        self.layers['LightNorm2'] = LightNormalization()
+        self.layers['Relu2'] = Relu()
+        self.layers['Conv3'] = DW_Convolution(self.params['W3'],
+                                           conv_param['stride'], conv_param['pad'])
+        self.layers['LightNorm3'] = LightNormalization()
+        self.layers['Relu3'] = Relu()
+        self.layers['Pool2'] = Pooling(pool_h=2, pool_w=2, stride=2)
+
+        self.layers['Conv4'] = Convolution(self.params['W4'],
+                                           1, 0)
+        self.layers['LightNorm4'] = LightNormalization()
+        self.layers['Relu4'] = Relu()
+        self.layers['Conv5'] = DW_Convolution(self.params['W5'],
+                                           conv_param['stride'], conv_param['pad'])
+        self.layers['LightNorm5'] = LightNormalization()
+        self.layers['Relu5'] = Relu()
+        self.layers['Pool3'] = Pooling(pool_h=2, pool_w=2, stride=2)
+
+        self.layers['Affine4'] = Affine(self.params['W6'])
+        self.layers['LightNorm6'] = LightNormalization()
+        self.layers['Relu6'] = Relu()
+
+        self.layers['Affine5'] = Affine(self.params['W7'])
+
+        self.last_layer = SoftmaxWithLoss()
+
+    def predict(self, x):
+        for layer in self.layers.values():
+            x = layer.forward(x)
+
+        return x
+
+    def loss(self, x, t):
+        y = self.predict(x)
+        return self.last_layer.forward(y, t)
+
+    def accuracy(self, x, t, batch_size=100):
+        if t.ndim != 1 : t = np.argmax(t, axis=1)
+        
+        acc = 0.0
+        
+        for i in range(int(x.shape[0] / batch_size)):
+            tx = x[i*batch_size:(i+1)*batch_size]
+            tt = t[i*batch_size:(i+1)*batch_size]
+            y = self.predict(tx)
+            y = np.argmax(y, axis=1)
+            acc += np.sum(y == tt) #numpy
+        
+        return acc / x.shape[0]
+
+    def gradient(self, x, t):
+
+        self.loss(x, t)
+
+        dout = 1
+        dout = self.last_layer.backward(dout)
+
+        layers = list(self.layers.values())
+        layers.reverse()
+        for layer in layers:
+            dout = layer.backward(dout)
+
+        grads = {}
+        grads['W1'] = self.layers['Conv1'].dW
+        grads['W2'] = self.layers['Conv2'].dW
+        grads['W3'] = self.layers['Conv3'].dW
+        grads['W4'] = self.layers['Conv4'].dW
+        grads['W5'] = self.layers['Conv5'].dW
+        grads['W6'] = self.layers['Affine4'].dW
+        grads['W7'] = self.layers['Affine5'].dW
+        return grads
+        
+    def save_params(self, file_name="params.pkl"):
+        params = {}
+        for key, val in self.params.items():
+            params[key] = val
+        with open(file_name, 'wb') as f:
+            pickle.dump(params, f)
+

+# coding: utf-8
+import sys, os
+sys.path.append(os.pardir)
+import time
+import numpy as np
+from dataset.cifar10 import load_cifar10
+from simple_convnet4 import SimpleConvNet
+from trainer import Trainer
+
+(x_train, t_train), (x_test, t_test) = load_cifar10(flatten=False)
+
+test_mask = np.random.choice(x_test.shape[0], 1000)
+x_test = x_test[test_mask]
+t_test = t_test[test_mask]
+
+max_epochs = 30
+
+network = SimpleConvNet(input_dim=(3,32,32),
+                        conv_param = {'filter_num': (32, 32, 64), 'filter_size': 3, 'pad': 1, 'stride': 1},
+                        hidden_size=512, output_size=10, weight_init_std=0.01)
+                        
+trainer = Trainer(network, x_train, t_train, x_test, t_test,
+                  epochs=max_epochs, mini_batch_size=100,
+                  optimizer='Adam', optimizer_param={'lr': 0.001},
+                  evaluate_sample_num_per_epoch=1000, early_stopping=5)
+start = time.time()
+trainer.train()
+elapsed_time = time.time() - start
+print ("elapsed_time:{0}".format(elapsed_time) + "[sec]")
+
+network.save_params("params.pkl")
+print("Saved Network Parameters!")
+
+markers = {'train': 'o', 'test': 's'}
+x = np.arange(trainer.current_epoch)
+plt.plot(x, trainer.train_acc_list, marker='o', label='train', markevery=2)
+plt.plot(x, trainer.test_acc_list, marker='s', label='test', markevery=2)
+plt.xlabel("epochs")
+plt.ylabel("accuracy")
+plt.ylim(0, 1.0)
+plt.legend(loc='lower right')
+plt.show()

+# coding: utf-8
+import sys, os
+sys.path.append(os.pardir)  # 親ディレクトリのファイルをインポートするための設定
+import numpy as np
+from optimizer import *
+
+class Trainer:
+    """ニューラルネットの訓練を行うクラス
+    """
+    def __init__(self, network, x_train, t_train, x_test, t_test,
+                 epochs=20, mini_batch_size=100,
+                 optimizer='SGD', optimizer_param={'lr':0.01}, 
+                 evaluate_sample_num_per_epoch=None, early_stopping=5, verbose=True):
+        self.network = network
+        self.verbose = verbose
+        self.x_train = x_train
+        self.t_train = t_train
+        self.x_test = x_test
+        self.t_test = t_test
+        self.epochs = epochs
+        self.batch_size = mini_batch_size
+        self.evaluate_sample_num_per_epoch = evaluate_sample_num_per_epoch
+
+        # optimzer
+        optimizer_class_dict = {'sgd':SGD, 'momentum':Momentum, 'nesterov':Nesterov,
+                                'adagrad':AdaGrad, 'rmsprpo':RMSprop, 'adam':Adam}
+        self.optimizer = optimizer_class_dict[optimizer.lower()](**optimizer_param)
+        self.early_stopping = EarlyStopping(patience=early_stopping, verbose=self.verbose)
+        
+        self.train_size = x_train.shape[0]
+        self.iter_per_epoch = max(self.train_size / mini_batch_size, 1)
+        self.max_iter = int(epochs * self.iter_per_epoch)
+        self.current_iter = 0
+        self.current_epoch = 0
+        
+        self.train_loss_list = []
+        self.train_acc_list = []
+        self.test_acc_list = []
+
+    def train_step(self):
+        early_stopping = False
+        batch_mask = np.random.choice(self.train_size, self.batch_size)
+        x_batch = self.x_train[batch_mask]
+        t_batch = self.t_train[batch_mask]
+        
+        grads = self.network.gradient(x_batch, t_batch)
+        self.optimizer.update(self.network.params, grads)
+        
+        loss = self.network.loss(x_batch, t_batch)
+        self.train_loss_list.append(loss)
+        if self.verbose: print(str(self.current_epoch) + " : " + str(int(self.current_iter % self.iter_per_epoch)) + " : train loss:" + str(loss))
+        
+        if self.current_iter % self.iter_per_epoch == 0:
+            self.current_epoch += 1
+            
+            x_train_sample, t_train_sample = self.x_train, self.t_train
+            x_test_sample, t_test_sample = self.x_test, self.t_test
+            if not self.evaluate_sample_num_per_epoch is None:
+                t = self.evaluate_sample_num_per_epoch
+                x_train_sample, t_train_sample = self.x_train[:t], self.t_train[:t]
+                x_test_sample, t_test_sample = self.x_test[:t], self.t_test[:t]
+                
+            train_acc = self.network.accuracy(x_train_sample, t_train_sample)
+            test_acc = self.network.accuracy(x_test_sample, t_test_sample)
+            self.train_acc_list.append(train_acc)
+            self.test_acc_list.append(test_acc)
+            early_stopping = self.early_stopping.validate(test_acc)
+
+            if self.verbose: print("=== epoch:" + str(self.current_epoch) + ", train acc:" + str(train_acc) + ", test acc:" + str(test_acc) + " ===")
+        self.current_iter += 1
+        return early_stopping
+
+    def train(self):
+        for i in range(self.max_iter):
+            if self.train_step():
+                break
+
+        test_acc = self.network.accuracy(self.x_test, self.t_test)
+
+        if self.verbose:
+            print("=============== Final Test Accuracy ===============")
+            print("test acc:" + str(test_acc))
+

+# coding: utf-8
+#import cupy as cp
+import numpy as cp
+import numpy as np
+
+def DW_im2col(input_data, filter_h, filter_w, stride=1, pad=0):
+    """다수의 이미지를 입력받아 2차원 배열로 변환한다(평탄화).
+
+    Parameters
+    ----------
+    input_data : 4차원 배열 형태의 입력 데이터(이미지 수, 채널 수, 높이, 너비)
+    filter_h : 필터의 높이
+    filter_w : 필터의 너비
+    stride : 스트라이드
+    pad : 패딩
+
+    Returns
+    -------
+    col : 2차원 배열
+    """
+    N, C, H, W = input_data.shape
+    out_h = (H + 2 * pad - filter_h) // stride + 1
+    out_w = (W + 2 * pad - filter_w) // stride + 1
+
+    img = np.pad(input_data, [(0, 0), (0, 0), (pad, pad), (pad, pad)], 'constant')
+    col = np.zeros((N, C, filter_h, filter_w, out_h, out_w))
+
+    for y in range(filter_h):
+        y_max = y + stride * out_h
+        for x in range(filter_w):
+            x_max = x + stride * out_w
+            col[:, :, y, x, :, :] = img[:, :, y:y_max:stride, x:x_max:stride]
+
+    col = col.transpose(1, 0, 4, 5, 2, 3).reshape(C, N * out_h * out_w, -1)
+    return col
+
+def smooth_curve(x):
+    """損失関数のグラフを滑らかにするために用いる
+
+    参考：http://glowingpython.blogspot.jp/2012/02/convolution-with-numpy.html
+    """
+    window_len = 11
+    s = np.r_[x[window_len-1:0:-1], x, x[-1:-window_len:-1]]
+    w = np.kaiser(window_len, 2)
+    y = np.convolve(w/w.sum(), s, mode='valid')
+    return y[5:len(y)-5]
+
+
+def shuffle_dataset(x, t):
+    """データセットのシャッフルを行う
+
+    Parameters
+    ----------
+    x : 訓練データ
+    t : 教師データ
+
+    Returns
+    -------
+    x, t : シャッフルを行った訓練データと教師データ
+    """
+    permutation = np.random.permutation(x.shape[0])
+    x = x[permutation,:] if x.ndim == 2 else x[permutation,:,:,:]
+    t = t[permutation]
+
+    return x, t
+
+def conv_output_size(input_size, filter_size, stride=1, pad=0):
+    return (input_size + 2*pad - filter_size) / stride + 1
+
+
+def im2col(input_data, filter_h, filter_w, stride=1, pad=0):
+    """
+
+    Parameters
+    ----------
+    input_data : (データ数, チャンネル, 高さ, 幅)の4次元配列からなる入力データ
+    filter_h : フィルターの高さ
+    filter_w : フィルターの幅
+    stride : ストライド
+    pad : パディング
+
+    Returns
+    -------
+    col : 2次元配列
+    """
+    N, C, H, W = input_data.shape
+    out_h = (H + 2*pad - filter_h)//stride + 1
+    out_w = (W + 2*pad - filter_w)//stride + 1
+
+    img = cp.pad(input_data, [(0,0), (0,0), (pad, pad), (pad, pad)], 'constant')
+    col = cp.zeros((N, C, filter_h, filter_w, out_h, out_w), dtype=np.float32)
+
+    for y in range(filter_h):
+        y_max = y + stride*out_h
+        for x in range(filter_w):
+            x_max = x + stride*out_w
+            col[:, :, y, x, :, :] = img[:, :, y:y_max:stride, x:x_max:stride]
+
+    col = col.transpose(0, 4, 5, 1, 2, 3).reshape(N*out_h*out_w, -1)
+    return col
+
+
+def col2im(col, input_shape, filter_h, filter_w, stride=1, pad=0):
+    """
+
+    Parameters
+    ----------
+    col :
+    input_shape : 入力データの形状（例：(10, 1, 28, 28)）
+    filter_h :
+    filter_w
+    stride
+    pad
+
+    Returns
+    -------
+
+    """
+    N, C, H, W = input_shape
+    out_h = (H + 2*pad - filter_h)//stride + 1
+    out_w = (W + 2*pad - filter_w)//stride + 1
+    col = col.reshape(N, out_h, out_w, C, filter_h, filter_w).transpose(0, 3, 4, 5, 1, 2)
+
+    img = cp.zeros((N, C, H + 2*pad + stride - 1, W + 2*pad + stride - 1), dtype=np.float32)
+    for y in range(filter_h):
+        y_max = y + stride*out_h
+        for x in range(filter_w):
+            x_max = x + stride*out_w
+            img[:, :, y:y_max:stride, x:x_max:stride] += col[:, :, y, x, :, :]
+
+    return img[:, :, pad:H + pad, pad:W + pad]