Hello friends. I’ve installed cygwin and python 2.7.13 within cygwin (running python 3 externally on Windows 10) and as I’m trying to run setup_p2.sh using the command bash setup_p2.sh I’m getting the following errors:
$ bash setup_p2.sh
: No such file or directory/bash
setup_p2.sh: line 4: $’\r’: command not found
setup_p2.sh: line 9: syntax error near unexpected token elif' 'etup_p2.sh: line 9:elif [ $region = “eu-west-1” ]; then
I’m running a dual boot with Ubuntu 16.04 on the other partition but I’m not at a point yet where I can run on my local machine. I haven’t seen anything about fixing this issue within Windows or using AWS through Ubuntu otherwise I wouldn’t be asking this seemingly introductory question. Any help resolving this would be appreciated.
I haven’t used Windows in a long while, but I remember that Cygwin was always a headache.
Have you given consideration to the option of installing a “Windows Subsystem for Linux”, which is possible in Windows 10? If I were you I would try that, it basically gives you a “true Linux” running alongside Windows.
Thanks for the quick reply. I’d actually much prefer to use Linux while connecting to the AWS servers. To do this do I just run setup_p2.sh after running install-gpu.sh and follow the same methodology for connecting via ssh? The only Ubuntu setup I’ve seen in my first look through the forum has to do with setting up my own server, which I may very well do after going through both parts of this but am not at the point of doing now.
I’m not sure I understand. Do you want to setup a local Linux environment from which to connect to another Linux environment in AWS? The setup_p2 script is intended for setting up your Linux environment in AWS. If you use AWS you don’t need a Linux on your own machine, you just need an ssh client.
I think you don’t understand because I’m not sure I fully understand. As far as I can tell, the set up video shows how to set up a connection to a GPU server on AWS in a Windows environment. Based on what you’re saying, using cygwin mimics a Linux environment on a Windows machine. If that’s the case then I just need to set up a connection from my local Linux machine to an AWS GPU cluster. Upon doing a bit more thinking, it seems to me that install-gpu is for creating a local server and all I need to do, after configuring AWS is simply run setup_p2.sh in my local Linux environment and it’ll set up the connection to AWS, yes?
Actually I am not the best person to help you with this, because I never used AWS – I set up my own Linux box from the start. But as far as I can tell, you run setup_p2.sh locally, and install-gpu.sh on your AWS instance.
I spent a fair bit on the machine, since I knew that I wanted to do a lot of deep learning experimentation, regardless of how this course went. Much of the budget went to the GPU. Here is the list of parts that I ordered, although oddly the total $ is higher than what I paid 3 months ago.
It has been very worthwhile for me, since it’s an excellent system and I’m using it all the time. Whether it is worthwhile for you depends on how serious you are about the subject, how much money you can spare, and how much free time you have (I didn’t want to get a weak machine, only to spend time upgrading a few months later).
All fair points. This might not be the place to ask this question, and I can understand if you don’t want to get into it, but how does deep learning differ from regular ML in a practical sense? I’ve done some basic ML modeling using sklearn (basic single models as well as ensemble stacking) to predict things like NFL spreads, NBA spreads, a basic fundamental stock algo, etc, and what I’ve found is that a large part of it is feature engineering and getting the right data. I know it’s something that I’d learn over the course of taking this course, but what sort of practical advantages would deep learning bring to the fold that would make spending the money worth it?