tensorflow张量流代码_实现简单的三维卷积训练、可视化

代码：

import numpy as np
import h5py
import tensorflow as tf
import time
from sklearn import preprocessing
import matplotlib.pyplot as plt
import math

# initial weight
def weight_variable(shape):
    initial = tf.truncated_normal(shape, stddev=0.1)  # 初始化赋值,符合正态分布,标准差为0.1
    return tf.Variable(initial)

# initial bias
def bias_variable(shape):
    initial = tf.constant(0.1, shape=shape) # 初始化bias,用0.1填充
    return tf.Variable(initial)

# convolution layer
def conv3d(x, W):
    return tf.nn.conv3d(input=x, filter=W, strides=[1, 1, 1, 1, 1], padding='VALID')

# pooling layer
def max_pool_3d(x):
    return tf.nn.max_pool3d(input=x, ksize=[1, 2, 2, 2, 1], strides=[1, 2, 2, 2, 1], padding='SAME')

# load data
cube_f_train = h5py.File("./data/cube_train_non_1159.hdf5", "r")
cube_f_test = h5py.File("./data/cube_test_non_1159.hdf5", "r")

cube_train_data = cube_f_train['data']

train_labels = cube_f_train['labels']

scaler = preprocessing.StandardScaler().fit(train_labels)
train_labels_n = scaler.transform(train_labels)

cube_test_data = cube_f_test['data']

test_labels = cube_f_test['labels']

test_labels_n = scaler.transform(test_labels)

# batch data
def get_batch_data(data, batch, batch_size):
    if data == 'train':
        batch_cube_xs = cube_train_data[batch * batch_size: (batch + 1) * batch_size]
        batch_ys = train_labe