virtualbox ubuntu 上的 SSH 服务器
我是 SSH 新手,有一些有关连接它的问题。 因此,我最近获得了访问大学 GPU 服务器的权限,并想知道如何利用它来加速我的强化学习训练,我主要使用 pytorch 和 openai 健身房环境以及稳定的基线。
在此之前,我在 virtualbox 上运行 ubuntu,并将所有依赖项下载到虚拟机中的 conda 环境中,但是当我连接到 SSH 服务器时,我无法再访问文件和 conda env。我后来了解到是因为我正在使用我连接的服务器中的文件系统。 是否可以在访问我的本地文件系统时利用 GPU 服务器,或者最好在 GPU 服务器内重新安装 conda 环境?另外,关于 python 脚本和环境(例如 mujoco),我是否必须在服务器上再次安装它?
我使用 ubuntu 的原因是因为我在使用 Windows 操作系统安装 mujoco 时遇到问题。我只能使用 Linux 操作系统来做到这一点,所以如果我可以不使用 virtualbox,而只连接到 GPU 服务器并访问 mujoco 环境,那就更好了,因为使用 virtualbox 有点滞后。 谢谢!
I am new to SSH and have a few questions regarding connecting to it.
So i was recently given access to the gpu server in the university and was wondering how to utilize it to speed up my reinforcement learning training, i mainly uses pytorch with openai gym environment and also stable baselines.
Before all this, i was running ubuntu on virtualbox with all my dependencies downloaded on conda environments in the VM, however when i connect to the SSH server i can no longer access the files and conda env. Which i later learnt was because i was using the filesystem in the server that i connected to.
Would it possible to utilize the gpu server while accessing my local file systems or would it better to re-install the conda environment inside the gpu server? Also with regards to the python scripts and environments such as mujoco, do i have to install it again it the server?
The reason why i use ubuntu was because i had issues with intalling mujoco using windows OS. I was only able to do so using linux OS, so if i can do away with using virtualbox and just connect to the GPU server and access the mujoco environments, that would be better as using virtualbox is kind of laggy.
Thanks!
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)
GPU Server通常已经设置了conda环境。
只需 conda 配置您的 python 版本环境并激活它即可。
喜欢:
conda create --name python34 python=3.4
source activate python34
以使用您的 Python 环境。或者
退出时
source deactivate python34
。GPU Server usually had set conda environment.
Just conda configure your python version environment and active it.
like:
conda create --name python34 python=3.4
source activate python34
to use your Python environment.or
source deactivate python34
when quit.