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

81–90 of 94 posts

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

#81

Post by ralphc »

Cacti wrote: post #52
cakedoggie wrote: post #37

Earlier quoted context omitted.

Why? Why can't I just do it with my crappy hardware, and it take longer?

You can do it on a poorer graphics card, but that's only if it has enough GPU RAM, and only if its supports a relatively recent compute version. You _can_ run it on a CPU, and that can work fine for very simple models (like the Tensorflow demos with MNIST) but anything much larger and things really start to take forever. Like, you'll be waiting for months.

I'm giving serious thought to buying or building a rig for these courses. What's "enough GPU RAM"?

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

#82

Post by jph00 »

edanm wrote: post #68
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…

Also the python distribution in question (anaconda) is becoming standard for data science - pytorch, for instance, (covered in part 2) assumes you're using anaconda for its default install method.

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

#83

Post by jph00 »

learner101 wrote: post #79
jph00 wrote: post #34

Earlier quoted context omitted.

We (including the rest of the learning community) would love to help you - could you post on the forums the problems you are having and we'll try to help sort them out? If you've already posted, please at-mention me there so I don't miss it. It's a fine line between helping too much vs too little in lesson 1! For the next time we run part 1, we're thinking we'll have an optional weekend workshop to teach the necessar…

Thanks for the reply. I had not (yet) posted on the forum as I was methodically searching through the previous replies to ensure my question had not already been answered. Thanks for crestle recommendation. I will now look into it. This should simplify my tool setup. If I still have questions around tool setup after posting, then I will at-mention you. I think the mini-MOOC (for people like me) is a great idea. I get…

Yes it'll be online a couple of months after completion, or you can apply to join remotely as an international fellow.

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

#84

Post by alexcnwy »

Honestly can't recommend this course highly enough.

It's definitely not perfect - the notebooks are not commented and the material does tend to jump around a bit - but what it does do, it does extremely well.

This course will teach you how to actually build deep learning systems and build the kinds of things you read about PhDs doing...

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

#85

Post by alexcnwy »

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.

If your data is in S3 then data transfer fees can suck but def cheaper than $ 0.90 per hour if you're regularly leaving models to train overnight :)

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

#86

Post by jph00 »

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…

As others have mentioned, installing anaconda is fine.

The question volume improves after lesson 3 when we get an audience mic.

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

#87

Post by alexcnwy »

jmeyers44 wrote: post #75
mcintyre1994 wrote: post #74

I've been looking to do part 1, so this is really cool - looking forward to this too! On http://course.fast.ai/part2.html the thumbnail for lesson 8 has specs for building a PC, with advice to use pcpartpicker. For part 1 I liked the idea of using AWS and only paying for a few hours, does part 2 have a hard requirement of a >$100s investment in hardware?

There is no requirement, but if you are spending a meaningful amount of money per month on AWS (over $100) and plan on working on DL projects for the next year or two it might make sense to make the initial up front investment.

You can definitely get by just using AWS and the flexibility is great but it can get expensive.

The Tiramisu model from lesson 14 takes like 25 hours to train and at $0.90 per hour, it adds up pretty fast...

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

#88

Post by Eridrus »

ralphc wrote: post #80
metafunctor wrote: post #33

Earlier quoted context omitted.

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.

Can this be done in Windows or is it better to install a Linux?

Better to install linux. Not all tools support Windows, and for those that do it is still a second class citizen and kind of annoying, e.g. look at installing CUDA/cuDNN on Windows vs Linux.

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

#89

Post by bradleyjg »

ralphc wrote: post #81
Cacti wrote: post #52

Earlier quoted context omitted.

You can do it on a poorer graphics card, but that's only if it has enough GPU RAM, and only if its supports a relatively recent compute version. You _can_ run it on a CPU, and that can work fine for very simple models (like the Tensorflow demos with MNIST) but anything much larger and things really start to take forever. Like, you'll be waiting for months.

I'm giving serious thought to buying or building a rig for these courses. What's "enough GPU RAM"?

A GTX 1060 with 6 GB of RAM is a good intro card. You can find them for The AWS P2 instances recommended for the course ($0.90/hour) have half a K80, which has access to 12 GB of VRAM. The list price for that card is $5000.

The GTX 1080 Ti has almost as much VRAM as half a K80, 11GB, and can be found for around $700.

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

#90

Post by ralphc »

bradleyjg wrote: post #89
ralphc wrote: post #81

Earlier quoted context omitted.

I'm giving serious thought to buying or building a rig for these courses. What's "enough GPU RAM"?

A GTX 1060 with 6 GB of RAM is a good intro card. You can find them for The AWS P2 instances recommended for the course ($0.90/hour) have half a K80, which has access to 12 GB of VRAM. The list price for that card is $5000. The GTX 1080 Ti has almost as much VRAM as half a K80, 11GB, and can be found for around $700.

I'm finding Acer rigs with GTX 1070, 8GB RAM. I'll probably get one of those, thanks.
Post reply on HN