파이썬과 다른 언어의 계산값을 비교하기 위한 파이썬 계산 코드 - LOPES-HUFS/DeepLearningFromForR GitHub Wiki

import numpy as np 
import tensorflow as tf
from tensorflow import keras
from keras.utils import np_utils

mnist = keras.datasets.mnist

(X_train, Y_train), (X_validation, Y_validation) = mnist.load_data()

X_train = X_train.reshape(X_train.shape[0], 784).astype('float64') / 255
X_validation = X_validation.reshape(X_validation.shape[0], 784).astype('float64') / 255

Y_train = np_utils.to_categorical(Y_train, 10)
Y_validation = np_utils.to_categorical(Y_validation, 10)

X_train.shape, Y_train.shape

x_100 = X_train[0:100]

x_100.shape

t_100 = Y_train[0:100]

t_100.shape

class TwoLayerNet:

    def __init__(self, input_size, hidden_size, output_size, weight_init_std=0.01):
        # 가중치 초기화
        self.params = {}
        self.params['W1'] = weight_init_std * np.random.randn(input_size, hidden_size)
        self.params['b1'] = np.zeros(hidden_size)
        self.params['W2'] = weight_init_std * np.random.randn(hidden_size, output_size)
        self.params['b2'] = np.zeros(output_size)

net = TwoLayerNet(input_size=784, hidden_size=100, output_size=10)

net.params['W1'].shape, net.params['b1'].shape, net.params['W2'].shape, net.params['b2'].shape

net.params['W1']

import h5py

hf = h5py.File('twoLayerNet.h5', 'w')

hf.create_dataset('W1',data=net.params['W1'], dtype=np.dtype('float64'))
hf.create_dataset('W2',data=net.params['W2'], dtype=np.dtype('float64'))
hf.create_dataset('b1',data=net.params['b1'], dtype=np.dtype('float64'))
hf.create_dataset('b2',data=net.params['b2'], dtype=np.dtype('float64'))
hf.create_dataset('x',data=x_100, dtype=np.dtype('float64'))
hf.create_dataset('t',data=t_100, dtype=np.dtype('float64'))

hf.close()

import h5py

hf = h5py.File('twoLayerNet.h5', 'r')

hf.get('W1')

W1 = hf.get('W1')
W1

np.array(W1)

W2 = hf.get('W2')
W2

b1 = hf.get('b1')
b1

b2 = hf.get('b2')
b2

hf.close()

from common.functions import *
from common.gradient import numerical_gradient

class TwoLayerNet:

    def __init__(self, W1, W2, hidden_size, output_size):
        # 가중치 초기화
        self.params = {}
        self.params['W1'] = np.array(W1)
        self.params['b1'] = np.zeros(hidden_size)
        self.params['W2'] = np.array(W2)
        self.params['b2'] = np.zeros(output_size)

    def predict(self, x):
        W1, W2 = self.params['W1'], self.params['W2']
        b1, b2 = self.params['b1'], self.params['b2']
    
        a1 = np.dot(x, W1) + b1
        z1 = sigmoid(a1)
        a2 = np.dot(z1, W2) + b2
        y = softmax(a2)
        
        return y
        
    # x : 입력 데이터, t : 정답 레이블
    def loss(self, x, t):
        y = self.predict(x)
        
        return cross_entropy_error(y, t)
    
    def accuracy(self, x, t):
        y = self.predict(x)
        y = np.argmax(y, axis=1)
        t = np.argmax(t, axis=1)
        
        accuracy = np.sum(y == t) / float(x.shape[0])
        return accuracy
        
    # x : 입력 데이터, t : 정답 레이블
    def numerical_gradient(self, x, t):
        loss_W = lambda W: self.loss(x, t)
        
        grads = {}
        grads['W1'] = numerical_gradient(loss_W, self.params['W1'])
        grads['b1'] = numerical_gradient(loss_W, self.params['b1'])
        grads['W2'] = numerical_gradient(loss_W, self.params['W2'])
        grads['b2'] = numerical_gradient(loss_W, self.params['b2'])
        
        return grads
        
    def gradient(self, x, t):
        W1, W2 = self.params['W1'], self.params['W2']
        b1, b2 = self.params['b1'], self.params['b2']
        grads = {}
        
        batch_num = x.shape[0]
        
        # forward
        a1 = np.dot(x, W1) + b1
        z1 = sigmoid(a1)
        a2 = np.dot(z1, W2) + b2
        y = softmax(a2)
        
        # backward
        dy = (y - t) / batch_num
        grads['W2'] = np.dot(z1.T, dy)
        grads['b2'] = np.sum(dy, axis=0)
        
        da1 = np.dot(dy, W2.T)
        dz1 = sigmoid_grad(a1) * da1
        grads['W1'] = np.dot(x.T, dz1)
        grads['b1'] = np.sum(dz1, axis=0)

        return grads

net = TwoLayerNet(W1=W1, W2=W2, hidden_size=100, output_size=10)

net.params['W1'].shape, net.params['b1'].shape, net.params['W2'].shape, net.params['b2'].shape

grads = net.numerical_gradient(x_100, t_100)

grads['b1']

hf = h5py.File('TESTtwoLayerNet.h5', 'w')

hf.create_dataset('W1',data=grads['W1'], dtype=np.dtype('float64'))

hf.create_dataset('W2',data=grads['W2'], dtype=np.dtype('float64'))

hf.create_dataset('b2',data=grads['b2'], dtype=np.dtype('float64'))

hf.create_dataset('b1',data=grads['b1'], dtype=np.dtype('float64'))

hf.close()