详解TensorFlow报”NotFoundError: Could not find requested filesystem type “的原因以及解决办法

背景

TensorFlow是深度学习领域的重要框架之一,但在使用过程中,可能会出现"NotFoundError: Could not find requested filesystem type "的错误。

错误描述

如下所示:

import tensorflow as tf

filename = '/data/test.tfrecord'

with tf.Session() as sess:
    # Read the contents of the file
    with tf.gfile.Open(filename, 'rb') as f:
        data = f.read()

报错信息

NotFoundError: Could not find requested filesystem type ''
; No error logs detected; check stderr log for any TensorFlow error logs. FileNotFoundError: No such file or directory

解决办法

方法一:升级tensorflow版本(推荐)

应该先尝试更新 TensorFlow 至最新的版本,看能否解决问题,具体升级方法如下:

pip install tensorflow --upgrade

方法二:指定文件系统类型

import tensorflow as tf

filename = '/data/test.tfrecord'
file_system = tf.gfile.GFile(filename, 'rb').filesystem

with tf.Session() as sess:
    # Read the contents of the file
    with tf.gfile.Open(filename, 'rb', file_system=file_system) as f:
        data = f.read()

方法三:显式使用文件系统

import tensorflow as tf

filename = '/data/test.tfrecord'

with tf.Session() as sess:
    # Read the contents of the file
    with tf.Graph().as_default(), tf.gfile.FastGFile(filename, 'rb') as f:
        data = f.read()

以上是解决TensorFlow报"NotFoundError: Could not find requested filesystem type "的原因以及解决办法的完整攻略,希望对大家的使用有所帮助。

此文章发布者为:Python技术站作者[metahuber],转载请注明出处:http://pythonjishu.com/tensorflow-error-43/

(0)
打赏 微信扫一扫 微信扫一扫 支付宝扫一扫 支付宝扫一扫
上一篇 3天前
下一篇 3天前

相关推荐