tensorflow--数据读取篇

最近，心血来潮搞一搞tensorflow,看着《tensorflow实战》码了几个简单的小网络，自以为蛮简单啊，当自己开始从头开始构建自己网络时候，就开始怀疑人生了。自己的数据读取都是一个大问题，今天解决了使用tensorflow读取csv文本数据，写到博客做个笔记。
`#!/usr/bin/env python2
# -- coding: utf-8 --
“””
Created on Tue Oct 17 08:24:47 2017

@author: wp
“””
import numpy as np
import pandas as pd
import tensorflow as tf

filename_queue = tf.train.string_input_producer([“feature/test.csv”])

reader = tf.TextLineReader()
key, value = reader.read(filename_queue)

record_defaults = [[0] for iter in range(2)]
features= tf.decode_csv(
value, record_defaults=record_defaults)
features = tf.to_float(features)

batch

batch_size = 1
mini_after_dequeue = 2
capacity = mini_after_dequeue+3*batch_size

example_batch,label_batch = tf.train