Ah, actually, the author's github repository has notebook, data, and additional tutorials. It's just that he did not link them in README.md. For example, you can see "Beginner MNIST" here ( https://github.com/Andrewnetwork/WorkshopScipy/blob/master/N... ) Since every collected project is periodically updated, the links in README.md might work sometime later.
I'm usually very careful when selecting a project, and his project gave me a hesitant moment. But, I thought the amount of material gone into building the project would outweigh inconvenience. I think I was short-sighted there. I'll make a note on this, and try to make clear communication with author in the future.
The examples are there to provide interesting implementations, tutorials, and/or articles as easy entries. I'd say some of them might look little dull to experienced veterans.
Nonetheless, they are valuable for those who seek to indulge into the fields. Hope you can look at them in that angle.
In terms of performance per cost, the cluster isn't on par with x86 system. Probably, not even close. Nonetheless, the network cost is real, and you get to have a first hand experience of how your algorithm performs.
It only serves two fronts; education and development.
Education-wise, you need something dirt-cheap to play with. For developer, you need some intermediate stage to deploy your work to test before going after a big cluster.
Yes I do. Since Apache Spark is chewing up memory like nothing, I cannot spare any portion of 1G mem for something else. This could lead to a shorter lifespan of SD cards though.
I've never imagined the RPI cluster to be taken seriously for production purpose. It only serves two purposes; education and developmental staging. For those two, you have a real hardware cluster that serves you superbly showing every nature you can experience with big-ass datacenter cluster.
In fact, storage is a big issue. RPI tends to destroy SD card in a way that a cluster like mine isn't exactly safe unless you put high-grade SD card and very reliable power source. Those contribute somewhat higher operation cost which violate the purpose of these single board computer.
See? this is funny. Some says this is the age of data. 10T is just a tiny amount of bits for some entities, and I cannot reliably transmit that much across pacific other than shipping actual HDD. :(
I'm trying to finish up a logger app for iPad. Based on NSLogger (https://github.com/fpillet/NSLogger), my goal is to build an app that could be useful out in the field when you're to debug or monitor how your app is behaving.