TensorFlow读取CSV数据

TensorFlow读取CSV数据

2017-05-12 10:46 by 猎手家园, 1723 阅读, 0 评论, 收藏, 编辑

代码来源于官方文档，做了一些小小的调整：

# -*- coding:utf-8 -*-
import tensorflow as tf

filename_queue = tf.train.string_input_producer(["file01.csv", "file02.csv"])
reader = tf.TextLineReader()
key, value = reader.read(filename_queue)

# Default values, in case of empty columns. Also specifies the type of the
# decoded result.
record_defaults = [[1], [1], [1]]
col1, col2, col3 = tf.decode_csv(value, record_defaults = record_defaults)
features = tf.stack([col1, col2])

init_op = tf.global_variables_initializer()
local_init_op = tf.local_variables_initializer()  # local variables like epoch_num, batch_size 可以不初始化local
with tf.Session() as sess:
    sess.run(init_op)
    sess.run(local_init_op)

    # Start populating the filename queue.
    coord = tf.train.Coordinator()
    threads = tf.train.start_queue_runners(coord=coord)

    for i in range(5):
        # Retrieve a single instance:
        example, label = sess.run([features, col3])
        print(example)
        print(label)

    coord.request_stop()
    coord.join(threads)

 

file01.csv 和 file02.csv 格式一样：

19,3,1
10,2,3
11,3,1
12,4,2
17,5,1
18,6,2
......