My tiny experiment playing with MNIST dataset of handwritten digits. Sadly, the end result can no longer be viewed live :( It guesses numbers pretty terribly, because my focus was more on how to setup the whole infrastructure and connect all the moving parts together. Backend is made using Flask and
TL;DR You should allow client to configure every direct output to him. At Oxipit a simple email notification feature produced 6 essential email settings and 7 extra settings regarding email content and attachments. Just had a quick exchange under the merge request between me and colleague Gediminas. The problem
One quote that buzzed my ear recently is by Jordan B Peterson where he said that the best way to learn to think is to learn to write. I remembered this dead blog, and thought it would be nice to make a return. In this post I will try to
In Linux your paths are case sensitive, if you have windows users who deal with case insensitive paths and want to communicate with them and handle their file requests you need to be able to "guess" which case is the correct one and return the file. This is
This caused some troubles for me, so I am sharing it -- to save someone else's precious time. Of course, solving problems is always rewarding in experience :) Scenario I want to write a JavaScript module, which would be using Ace Editor. I decide to do it using TDD,
Nice quote by Henry Ford about how people see innovations: If I had asked people what they wanted, they would have said faster horses. So don't listen to what words are being used, but try to understand what need hides behind them.
Recently I had to create a list of things one should know to better understand authors who are using LaTeX to write their article. Yes, LaTeX is not just commands, it is a fully functional programming environment so it has to be treated as such. These are some of the
This simple Python program example produces nice human readable word differences splitting at whitespaces, be it space, tab, or newline: from diff_match_patch import diff_match_patch dmp = diff_match_patch() c1 = 'some long\nfile with newlines' c2 = 'some long\nfiles\twith no newlines' tmp
Recent article in Delfi.lt provided some interesting ideas and numbers. There is one parameter, let's call it P, which shows how much of particular trait is influenced by environment, and how much by genes. Let's say if P=1 then the trait is totally caused
I am going to regret this, but I signed up for two additional classes by (ex) Stanford Professor Sebastian Thrun: * CS 373: PROGRAMMING A ROBOTIC CAR * CS 101: BUILDING A SEARCH ENGINE They both can be found at http://www.udacity.com. Watch Sebastian's talk about AI class,
After finishing AI and ML classes, it's time to look ahead and sign up for the new ones. The whole list of free lectures by professors of Stanford/Berkely/University of Michigan can be found in http://www.class-central.com. I decided to sign up for 5 of
If you simply write: <xsl:value-of select="//some-element"/> you get just a string of it's internal text elements, no node names. It's naked. If you write: <xsl:copy-of select="//some-element"/> You get just the copy of everything. You