AWS SageMaker Studio Lab-允许拒绝
我想使用AWS SageMaker Studio Lab来训练一些深度学习模型,但是当我想使用火车时,py Sagemaker会返回有关stdnfile的错误。
oserror:找不到sndfile库
之后,我发现我可以使用
apt -get install libsndfile1 -y
来解决该错误,但是每次尝试时,我都会得到一个下面的错误。
e:列表目录/var/lib/apt/lists/partial缺少。 - 获取(13:拒绝许可)
我尝试使用sudo
,因为AWS SageMaker使用基于Linux的发行版,但它返回
bash:bash:sudo:under condery
>
我尝试了su
,它想要我的密码。我尝试了我的帐户密码,但它拒绝了。
顺便说一句,我在的不同问题中看到了一个答案 - 允许 - 根
,但它也没有起作用。
PS:我在AWS SageMaker Studio中使用了相同的Python文件和apt -get install libsndfile1 -y
,它运行良好。 AWS SageMaker服务器类型为 m5.24xlarge 和 ml.t3.medium (CPU服务器)。我在AWS SageMaker Studio中要求GPU服务器,但我仍在等待( ml.g4dn.xlarge )。另一方面,我正在尝试在AWS Sagemaker Studio Lab中与GPU进行培训。
I want to use AWS SageMaker Studio Lab for training some deep learning models but when I want to use my train.py SageMaker returns an error about stdnfile.
OSError: sndfile library not found
After that I found out I can fix that error with
apt-get install libsndfile1 -y
but every time when I try, I'm getting an error in the below.
E: List directory /var/lib/apt/lists/partial is missing. - Acquire (13: Permission denied)
I tried to use sudo
because AWS SageMaker uses a Linux-based distro but It returns
bash: sudo: command not found
I tried su
either and It wants a password from me. I tried my account password but it rejects.
Btw I saw an answer in the different question for --allow-root
but It didn't work either.
p.s.: I use the same python file and apt-get install libsndfile1 -y
in AWS SageMaker Studio and It worked well. AWS SageMaker servers type were ml.m5.24xlarge and ml.t3.medium (CPU Servers). I request a GPU Server in AWS SageMaker Studio but I'm still waiting (ml.g4dn.xlarge). On the other hand, I'm trying to train with GPU in AWS SageMaker Studio Lab.
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
Studio Lab尚不支持APT安装,请参见 github问题#127 。
正如您提到的,SageMaker Studio允许安装 - 您应该能够获得快速启动GPU实例,请重试创建该应用程序。另外,请确保您的帐户在Studio上有足够的限制。
apt installs aren't supported yet for Studio Lab, see GitHub issue #127.
SageMaker Studio, as you mentioned, allows installs - you should be able to get a fast launch GPU instance, please retry creating the app. Also, make sure your account has sufficient limits for GPU notebooks on Studio.