Compiling TensorFlow with GPU support on a MacBook Pro

OK, so TensorFlow is the popular new computational framework from Google everyone is raving about (check out this year’s TensorFlow Dev Summit video presentations explaining its cool features). Of course, a fun way to learn TensorFlow is to play with it on your own laptop, so that you can iterate quickly and work offline (perhapse build a hot dog recognition app). In these cases a GPU is very useful for training models more quickly. There used to be a tensorflow-gpu package that you could install in a snap on MacBook Pros with NVIDIA GPUs, but unfortunately it’s no longer supported these days due to some driver issues. Luckily, it’s still possible to manually compile TensorFlow with NVIDIA GPU support. I’ve hunted through a lot of different tutorials (1, 2, 3, 4 – this last one helped me the most) to bring you this hopefully complete description of how to set everything up correctly and get deep into learning (and I know, in 2 months probably become just another one in that list of outdated tutorials, but that’s life 🙂 ).

For the sake of verbosity, I’m using a MacBook Pro 10,1 with an NVIDIA GT 650M and OS X 10.12. Hopefully, though, it will work on a couple of other configurations as well. In any case, let’s start… Continue reading Compiling TensorFlow with GPU support on a MacBook Pro

Simple way to set up Django and a SPA frontend on Heroku

This post was adapted from my Stack Overflow answer to a question about using single page applications (SPAs) with Django on Heroku.

Update: check out my django-spa package for a ready-to-use solution for serving SPAs from Django.

Here’s how to set up Django to serve your static files and index.html on / while still having the possibility to use Django views for the admin dashboard, registration etc.:

from django.conf.urls import include, url
from django.contrib import admin
from django.contrib.staticfiles.views import serve
from django.views.generic import RedirectView

admin.autodiscover()

urlpatterns = [

    # / routes to index.html
    url(r'^$', serve,
        kwargs={'path': 'index.html'}),

    # static files (*.css, *.js, *.jpg etc.) served on /
    # (assuming Django uses /static/ and /media/ for static/media urls)
    url(r'^(?!/?static/)(?!/?media/)(?P<path>.*\..*)$',
        RedirectView.as_view(url='/static/%(path)s', permanent=False)),

    # other views still work too
    url(r'^admin/', include(admin.site.urls)),
]

Continue reading Simple way to set up Django and a SPA frontend on Heroku

Disk on a platter

This article on building a Raspberry Pi NAS solution has been sitting in my drafts for three years now. Recently, my Raspberry Pi had stopped working (the SD card had died), so I rebuilt it yesterday. I cursed myself for not having finishing the text back then, as I now had to retrace some steps manually. So, here goes the finished procedure for future reference.

I got a Raspberry Pi as a birthday present from my thoughtful colleagues!

Raspberry Pi

Now, as my first project I decided to connect it to my 2TB external hard drive and serve it on my local network using the Samba protocol (I tried NFS and SSH too, but Samba proved to be the most performant protocol and is also cross-platform). No more moving the disk around and hooking up USB cables 🙂 Continue reading Disk on a platter

CloudFleet – Captain’s log – The End of the Sysadmin

We have already talked quite a bit about the importance of self-hosting your apps due to the limitations of centralized cloud services and mentioned a number of great apps ready for self-hosting. In this blog post, we assume that you already self-host your apps to protect your data autonomy. From here, we look at what happens after you set up an app on your hardware, and how you maintain it: a job that is typically done in organisations by professional system administrators, aka sysadmins.

Source: CloudFleet – Captain’s log – The End of the Sysadmin

Home automation using Python, Flask & Celery

We are all a bit lazy in this post-holiday period, so what better project to work on during these relaxed evenings at home, but on a home automation system. Having Docker containers on a physical device that has access to all other IoT devices in our network with exposed APIs like TVs, speakers or maybe even droids and being able to iteratively upgrade these containers gives us ample opportunity to play.

I love the elegance of resin.io’s Docker container deployment & upgrade method, so I use it a lot for hobby projects & freelance work. In this tutorial, I’ll show you how to create a Python Flask app with periodic Celery tasks for controlling your TV via the Chromecast API. All of the source code can be found in this repo. So, go get a hot cup of tea, clone the repo and let’s get started…

Source: Home automation using Python, Flask & Celery