Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Deep Learning for Coders – Launching Deep Learning Part 2

fast.ai

Post reply on HN
Print view View original

61–70 of 94 posts

Re: Deep Learning for Coders – Launching Deep Learning Part 2

#61

Post by cbrevik »

yorwba wrote: post #59
cbrevik wrote: post #57

Earlier quoted context omitted.

I'm doing part 1 at the moment, and the outdated AWS setup instructions on the site frustrated me a bit as well. Maybe an idea would be to leverage the community a bit more? Make it easier to edit the wiki for MOOC-students which are the ones watching/reading the material later on. And maybe refer to that wiki more on the site. I end up using the forums a lot for trying to find info on stuff that is not working, or I…

It's been a while since I watched the course, but IIRC the setup video had an overlay that said to simply put "fast.ai MOOC" (or something similar) into the P2 request form. Worked perfectly for me. Did that not show up for you?

Yeah, that showed up for me. It was more about how to use the setup scripts, they had changed a bit it seemed.

I thought if the resources around the course evolve over time, then the material should as well. It would be easier to keep a wiki up-to-date and refer to that, instead of a video (and site). For setup and "sidenotes" specifically, I'm not saying the lessons should be re-created in the wikis.

Although the homework and lesson notes in the wikis was indispensable.

Re: Deep Learning for Coders – Launching Deep Learning Part 2

#62

Post by praseodym »

mlboss wrote: post #54
adam12 wrote: post #32

Earlier quoted context omitted.

I started watching the videos in Part 1 and was disappointed to find out that I was going to have to create an AWS account. I really don't want to deal with accidentally racking up a large bill on AWS.

For GPU instance, I would suggest https://www.hetzner.com/dedicated-rootserver/ex51-ssd-gpu?co... . It is only 99 euro/month.

The 99 euro setup fee is a dealbreaker if you'd want to use it just for some courses.

Re: Deep Learning for Coders – Launching Deep Learning Part 2

#63

Post by acidburnNSA »

metafunctor wrote: post #33
adam12 wrote: post #32

Earlier quoted context omitted.

I started watching the videos in Part 1 and was disappointed to find out that I was going to have to create an AWS account. I really don't want to deal with accidentally racking up a large bill on AWS.

If you have the hardware, you can do without an AWS account (you might need to be an "expert" to interpret the instructions to run on your own hardware instead of AWS). I used my gaming rig with a 1080 GTX Ti card.

I've been trying to get it up and running using the latest versions of the libs. I have theanos finding my GPU now and running basic cuda code but cannot get the cuDNN libraries to be recognized even after installing a few versions of the Nvidia site. Ugh. Giving up for now.

Re: Deep Learning for Coders – Launching Deep Learning Part 2

#64

Post by natch »

Afraid I may have missed the window on the chance to provide feedback to jph00 via this channel, but here goes.

Am watching Part 1 now and only two sessions in, but there are some tweaks I would love to see. First the positive: I really appreciate the approach of hands-on and teaching theory only as it's needed and in conjunction with applied work.

Would love to see a tiny bit of time spent on setting up tools for people who already have good Nvidia GPU systems. My Ubuntu system has python (2.7) and python 3.5 both installed, but no Anaconda... I don't know if I'm going to totally screw up my system if I install Anaconda over those working existing python installations, for example.

It would be great to hear the questions. I can barely hear a faint voice in the background as Rachel reads the questions (presumably from online) but it seems like it would be a very easy tweak to have her closer to a microphone. Maybe this happens in later sessions and I just haven't gotten to them yet.

It would be great if so many things weren't abbreviated in the code variable and function names. Examples: nb for notebook, t for ?, a for array(?), U, s, and Vh for ?, ims (?), interp (interpretation or interpreter or interpolation?), sp, v, r, f, k, trn (train or turn or something else?), pred (predicate or prediction?), vec_numba (?)... the list goes on. Yes if I knew the field these might be obvious but for some of them I'm still learning. "np" I understand since that's standard practice and you explained it. It would be really really easy to just spell out words in the code, as well as being a good practice in general imho, and, since you are trying to teach stuff, it would seem appropriate.

Those nitpicks aside I'm really stoked about the course and really appreciate everything you have been putting into it!

Re: Deep Learning for Coders – Launching Deep Learning Part 2

#65

Post by mkl »

natch wrote: post #64

Afraid I may have missed the window on the chance to provide feedback to jph00 via this channel, but here goes. Am watching Part 1 now and only two sessions in, but there are some tweaks I would love to see. First the positive: I really appreciate the approach of hands-on and teaching theory only as it's needed and in conjunction with applied work. Would love to see a tiny bit of time spent on setting up tools for pe…

> Would love to see a tiny bit of time spent on setting up tools for people who already have good Nvidia GPU systems. My Ubuntu system has python (2.7) and python 3.5 both installed, but no Anaconda... I don't know if I'm going to totally screw up my system if I install Anaconda over those working existing python installations, for example.

Anaconda lives in its own folder (usually in $HOME). You can't screw anything up by installing it, and in fact you can hardly tell it's there. You need to set your path to actually use Anaconda's programs, and you shouldn't do that in .bashrc, but just in the shells where you are actively using it, with something like:

  export PATH=/home//Development/Tools/Anaconda3/bin:$PATH

Re: Deep Learning for Coders – Launching Deep Learning Part 2

#66

Post by tminima »

natch wrote: post #64

Afraid I may have missed the window on the chance to provide feedback to jph00 via this channel, but here goes. Am watching Part 1 now and only two sessions in, but there are some tweaks I would love to see. First the positive: I really appreciate the approach of hands-on and teaching theory only as it's needed and in conjunction with applied work. Would love to see a tiny bit of time spent on setting up tools for pe…

I haven't started on the course yet, but I have exp. with Anaconda. On every fresh linux install, I install Anaconda first. It won't do anything to your system. It'll ask for a path, install all the necessary files there and then add that directory at the start of path so that when you type python on terminal the anaconda python will start. You can still access the previously installed python version by typing python2.7 or python3.5.

Re: Deep Learning for Coders – Launching Deep Learning Part 2

#67

Post by cbrevik »

colmvp wrote: post #39

My feelings on Part 1: I felt like the setup of the first part was at time a little frustrating, since I started it during a time when Keras had switched to a newer version which wasn't compatible some of the utility code that was written. Add this to the newbie factor to notebooks, and it was a pretty rough first week or so to setup and get actual learning done. It took me a bit of time to realize notebooks were mor…

Totally agree on the unstructured content in the forums. I would suggest they open up the wikis more for editing by MOOC-students, and work to keep that up-to-date. Wading through forums for those nuggets of information to fix your problems is time consuming and boring.

Pro-tip: Press CMD+F (or CTRL+F) a second time to "override" the shortcut for search.

Re: Deep Learning for Coders – Launching Deep Learning Part 2

#68

Post by edanm »

metafunctor wrote: post #24

Part 1 was great. However, the first lesson took a bit of stamina to go through. Much of it was introducing basic Unix/AWS/shell/Python things I know intimately and have strong opinions and deeply set ways about. Shell aliases, how to use AWS, what Python distribution to run, running Python from some crazy web tool called notebooks (and not Emacs), etc. felt like I was forced to learn a random selection of randomly f…

You're partially right. But I would like to mention that in many ways, their way is the "standard" way to do things.

Just as an example, the "crazy web tool called notebooks" is pretty much the standard way of working in many different areas of machine learning/deep learning (and others). It doesn't mean you necessarily have to do it yourself, but tons of material out there will be in this format, so it's valuable in and of itself to know it.

Also, you really should give notebooks a chance, they're a game-changer for productivity, IMO. (Although I'm a vim user, not emacs, so maybe you shouldn't listen to me ;).

Re: Deep Learning for Coders – Launching Deep Learning Part 2

#69

Post by ChristianGeek »

colmvp wrote: post #39

My feelings on Part 1: I felt like the setup of the first part was at time a little frustrating, since I started it during a time when Keras had switched to a newer version which wasn't compatible some of the utility code that was written. Add this to the newbie factor to notebooks, and it was a pretty rough first week or so to setup and get actual learning done. It took me a bit of time to realize notebooks were mor…

Which book?

Re: Deep Learning for Coders – Launching Deep Learning Part 2

#70

Post by ErikBjare »

praseodym wrote: post #60
anurag wrote: post #55

Earlier quoted context omitted.

(I built Crestle.) The downside is you don't get as much control over your environment (you get a terminal in the browser, but no SSH or sudo access, for example). If you have the inclination and ability to manage your own instances, EC2 is more flexible.

Just wondering: what kind of infrastructure are you running on? If I were to build something like this I'd probably run it atop of EC2 or GCE, so pricing would never be lower than them.

If people aren't fully utilizing their instances, you can offer lower prices since you can have more than one user per underlying instance.

Not sure that's how he does it, but it's one way.

Post reply on HN