Ladvien's Lab

After having success with training a CNN on our initial dataset, we decided to up the game on generating training images. My buddy Rockets built a nice little turntable and ordered a couple of NEMA17s for each of us. His idea was we could both start generating training images.

I asked if he would be OK with me ordering some RAMPs boards and programming them to synchronize with the PiCamera. I figured, it would probably be better for reproducibility if we had solid hardware, with custom firmware and software.

After a few hours of coding over a couple of weeks ...

This article is part of a series. It should explain the code used to train our convolutional neural-network (CNN) LEGO classifier.

If you want to code along with this article, we've made it available in Google's Colab:

• Lego Classifier

Or if you want to run the code locally:

• lego_sorter

It's a WIP, so comment below if you run into any issues.

Classifier Code:

Our code started with a notebook found on Kaggle:

• Lego Brick Images Keras CCN

However, there problems in the code. I rewrote most of it, so I'm not sure how much of the original is left. Still ...

I've a robot friend. To be clear, the friend is not a robot, rather, we build robots together. One of the projects we tossed about is building a LEGO sorting machine. Rockets is the friends name--again, not a robot--teaches robotics to kids. For their designs, LEGOs are the primary component. Unfortunately, this results in much time spent to preparing for an event.

He mentioned to me, "What I really need is a sorting machine." And proceeded to explain his plain for building one.

I was skeptical for some time, but finally, I got drawn in he talked about incorporating a ...

The last two articles have been getting oriented to SQL, however, the information in them will disappear quickly if we don't give you a way to practice on data meaningful to you. Let's face it, as much fun as it is to find out random employees salaries, those don't mean anything to you .

This article will show you how to setup a copy of MySQL Server on your PC, connect to it, load data from a CSV, and query those data. There's a lot to get done, so let's get started.

Local MySQL Server Setup

Each of the three operating ...

Welcome back! Alright, now we know how to connect to a remote server from within MySQL Workbench, let's start writing some queries.

Here's a common SQL query:

    SELECT e.emp_no,
           e.last_name,
           t.title
      FROM employees AS e
 LEFT JOIN titles    AS t
        ON e.emp_no = t.emp_no
     WHERE e.hire_date > '1999-12-31'
  ORDER BY e.last_name DESC;

This query produces the following table when run on our employees database.

I'm usually writing about hacking, robotics, or machine learning, but I thought I'd start journaling thoughts on data analytics, which is how I pay the bills these days. I wanted to begin with a series on MySQL, as I've some friends I feel it'd help enter the field. But, I'll eventually expand the series to include visualizations, analysis, and maybe machine learning. And I hope these articles help someone move from manually generating reports in Excel to writing scripts that'll automate the boring stuff. As I like to say, "knowing to code gives you data superpowers!"

I'm a professional data ...

We're almost done. In the previous articles we've used a local machine to train a CNN to detect toxic sentiment in text. Also, we prepared a small (1GB RAM) server to use this pre-trained network to make predictions. Now, let's finish it and create a webservice where anyone can access our awesome magical algorithm.

Prediction Service

On your remote server, navigate to your flask_app folder and create a file called nn_service.py . The following code creates an HTTP request endpoint /detect-toxic and it exposes to other programs running on the server. A bit more explanation after the code.

cd /home ...

Previously, I wrote about training a CNN to detect toxic comments from text alone. But, I realized, even if one has a nice little NN to solve all the world's problems it doesn't help unless it is in production.

This article is going to cover how to prepare a server and needed word embeddings to mechanize the NN in a Flask webservice.

Server Setup: Preamble

For this project I'm using a small server from Linode--called a "Nanode." At the time of writing these servers are only $5 a month. The catch? They only have 1GB of RAM. It's definitely going ...

I'm writing learning-notes from implementing a "toxic comment" detector using a convolutional neural network (CNN). This is a common project across the interwebs, however, the articles I've seen on the matter leave a few bits out. So, I'm attempting to augment public knowledge--not write a comprehensive tutorial.

A common omission is what the data look like as they travel through pre-processing. I'll try to show how the data look before falling into the neural-net black-hole. However, I'll stop short before reviewing the CNN setup, as this is explained much better elsewhere. Though, I've put all the original code, relevant project ...