It will tell you if your hives are making honey

It all started several years ago with an old dream: to have my own hives and make my own honey.

I took training (it’s a trade), and had 3 hives for a couple of years. My record is 50kg of honey with one hive. Unfortunately, last spring, with COVID, I could not take good care of them and I lost two.

This year, I decided to reboot this project, with a small plus: a connected bee hives.

I know there are several projects out there, even commercial offerings but it…


Figure out from your coach, with a DIY solution.

It all started with a life or death question, how much hot water do we have for showers ?

I know very few people ever wonder, they take it for granted.


It all started with a user need and it is related to one of the key elements of any household: the fridge.

Sometimes my fridge’s door does not close completely. If one pushes it too hard, it rebounds and stays slightly open. Not open enough to be noticed, just a few millimeters, but open enough so that, in the morning, the fridge’s content is a bit damp. Besides, it is not very energy efficient.

For a while, I just ignored the problem. “It will go away” .

But one day, I decided to attack it with DIY electronics and some…


I do not know how to play music. But I can still play with music.

This is part of a series of articles, which explore many aspects of this project, including static MIDI file generation, real time streaming, Tensorflow/Keras sequential and functional models, LSTM, over and under fitting, attention, embedding layers, multi-head models, probability distribution, conversion to TensorflowLite, use of TPU/hardware accelerator, running inferences on Raspberry PI, ….

See Part 1 and Part 2.

What could go wrong during training ?

In Part 2 we looked at how validation loss evolves as the network gets trained.

But your first neural network training may very well look…


I do not know how to play music. But I can still play with music.

This is part of a series of articles, which will explore many aspects of this project, including static MIDI file generation, real time streaming, Tensorflow/Keras sequential and functional models, LSTM, over and under fitting, attention mechanism, embedding layers, multi-head models, probability distribution, conversion to TensorflowLite, use of TPU/hardware accelerator, running application on multiple platforms (Raspberry PI, edge devices) ….

See Part 1 . The entire code is available on github

Let’s go thru the process of preparing data and training our first neural network.

Trained…


I do not know how to play music. But I still can play with music.

source twinlk.com

It all started with a long time regret. My parent did not offer me music lessons when I was a kid.

So I do not know how to play music, but I still can play with music.

How? My idea is to use deep learning to automatically generate music. This music will be inspired by a given author and hopefully will sound interesting.

Even if the generated music will resemble the music of the actual composer, it will be based on my personal tuning of…


and do it from your coach

Deep blue in the mountains

It all started with an engineer guilty thought. Let’s fix something which is working fine.

A while ago, I had developed a string lights which automatically blinks at night. The system was rather primitive, based on the Arduino nano board, and built with a few lines of C code.

It was powered from a wall plug, and hung from my garage.

I thought: for Christmas, let’s make it better, solar powered, connected to my home network, and controlled from a smartphone. And I rewrote it in MicroPython (a version of Python for microcontroller).

Actually…


It all started with a snow problem.

I wanted to check remotely how much snow had fallen on my home in the mountains (I live in the French Alps), specially whether the snow removal machine had cleared the way.

So I started to tinker with the camera module of the Raspberry Pi (a low cost, Linux based, single board computer).

I quickly landed on Motion , a software package which provides real time camera feed into any web browser. …


It runs with the Sun

It all started with the word ‘solar’.

As I was tinkering with the Raspberry PI, I latched on PiJuice, a Kickstarter project promising solar power for the PI.

PiJuice is a Raspberry PI add-on board (aka HAT — Hardware Attached on Top) which includes a LIPO battery, a RTC (real time clock), a UPS (uninterrupted power supply), and a software environment to program the startup and safe shutdown of the PI.

The LIPO battery can be charged from USB or from a solar panel. I picked a 6 Watt solar panel.

The PIJuice HAT on top of a Raspberry Zero.
The PIJuice HAT on top of a Raspberry Zero.
The PiJuice Hat on top of a Raspberry Zero. The Solar panel is connected to the PiJuice via micro-USB. A USB Webcam is connected to the Raspberry. The Blue led indicates the LIPO is charging. The enclosure is a simple electricity junction box.

I quickly realized that…


Build it, it’s affordable. Move it, it’s battery powered.

It all started with My Love sending me a Medium article.

She is half French and half American (not sure which half is bigger), and an avid reader of the New York Times. This article caught her eyes.

I went thru this article , and this other one titled “ This Super-Clean Smart Screen Puts a Newspaper on Your Wall“ and was intrigued by the application, and seduced by the concept behind it : Calm technology.

Both articles are about displaying, every morning, the front page of the New York Times…

pascal boudalier

Tinkering with Raspberry PI, ESP32, Solar, LifePo4, mppt, IoT, Zwave, energy harvesting, Python, MicroPython, Keras, Tensorflow, tflite, TPU. Ex Intel and HP

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store