def getConv1DOutputSize(inputSize, kernelSize, stride, padding): output = int((inputSize + 2 * padding - kernelSize) / stride + 1) return output def getConv2DOutputSize(inputSize, kernelSize, stride, padding): output = [getConv1DOutputSize(inputSize[0], kernelSize, stride, padding), getConv1DOutputSize(inputSize[1], kernelSize, stride, padding)] return output def getConvT1DOutputSize(inputSize, ..

import numpy as npimport scipy.stats as stats def KLD(pk, qk): kld = stats.entropy(pk, qk) print(kld) # ppk = np.random.normal(0.1, 0.01, 100000)qk = np.random.normal(0.1, 0.01, 100000)KLD(pk, qk) pk = np.random.normal(0.1, 0.001, 100000)qk = np.random.normal(0.1, 0.001, 100000)KLD(pk, qk) 출처 : https://m.blog.naver.com/PostView.nhn?blogId=atelierjpro&logNo=220981354861&proxyReferer=&proxyReferer..

import torchimport torch.nn as nnimport torch.nn.functional as Fimport torch.autograd as autogradimport torch.optim as optimimport numpy as npfrom torch.autograd import Variablefrom torchvision import datasetsfrom torchvision import transformsfrom torchvision.utils import save_image '''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''# Mnist 데..

import tensorflow as tffrom tensorflow.examples.tutorials.mnist import input_data Mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)NUM_CLASSES = 10TRAIN_EPOCH = 10BATCH_SIZE = 100LEARNING_RATE = 0.001 # 0.1 : Overshooting x = tf.placeholder(tf.float32, [None, 28*28])y_ = tf.placeholder(tf.float32, [None, NUM_CLASSES]) # Convolutional Layerx_img = tf.reshape(x, [-1, 28, 28, 1])LayerC..