:【None,14,14,32】输出:【None,14,14,64】(64个过滤器 ,每层都是32个14*14相加)
激活:输出:【None,14,14,64】
池化:2*2 strides=2...(x_reshape, weight_conv1, strides=[1, 1, 1, 1], padding="SAME") + bias_conv1)
# 池化 2*2 strides2...strides=[1, 2, 2, 1], padding="SAME")
with tf.variable_scope("conv2"):
# 3,二层卷积 卷积5*5*32...]
x_relu2 = tf.nn.relu(tf.nn.conv2d(x_pool1, weight_conv2, strides=[1, 1, 1, 1], padding="SAME..., ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding="SAME")
# 4,全连接层
with tf.variable_scope("