sdlion

If I recall correctly, electrolytic capacitors have to get "burned in" during manufacturing to make them work.

A current is passed through the capacitor and a thin film of oxide is built up in one of the terminals, according to the polarity. This is why electrolytic caps have polarity, if you use them with their polarity inverted, you flake off that oxide layer and thus short them out.

A free running current in that electrolyte boils it off, and you get an exploding cap

As a note, for Japanese text deepl is widely used even by Japanese people. From eng to jpn it may not choose properly nuanced words though, but it largely produces acceptable translations.

In my experience, being a workplace or even academia, there's as many types of work environments as types of teachers.

In the end, management (or founders if it's a startup) set the tone and who they end up hiring (thus, who will manage your work). And management is just normal people in the end.

I guess because at server farm level, performance/efficiency translate to real million USD savings. In general, at scale ends (the cloud and the embedded) this matters a lot. In resource limited environments like raspberry pi, this design philosophy wins over many users between DIY and the private industry.

For training, more GPU RAM will allow you train with greater resolutions in less time and better performance. Before feeding data to a model, it needs to be resized to a "network dimension" (YOLOv4 default is 416x416 px if I recall correctly). For training, it will group several samples and train with them at the same time, in "batches". For better generalization you want bigger batches (so more different images are feed at the same time). With a 3060 (non-Ti) you'll have 12GB of GPU RAM, with that I think you can run the default settings (network size, batch size and subdivision of batches) for the YOLOv4 model. If you want to go to 512px, you might have to increase the subdivision (create more subbatches) or reduce the batch size. If I recall correctly, you could find 3070 with less than 12GB of RAM, so in trying to purse faster training times (I'm not talking about inference, using the model to actually recognize something) you might not be able to train with a broader range of options that can improve your accuracy.

It would seem that it depends on the architecture you will be using. Whether is a ARM, GPU, mobile GPU processor, etc. This is comment from the author of YOLOv4 mentions that NanoDet is more suitable for ARM-CPU's https://twitter.com/alexeyab84/status/1436377831974506496

Is not the same scale cut out used in Scaled Yolov4 and previous research? (Though this one is on AP not mAP). It seems it is needed since the improvement would be hard to appreciate if it would show a 0-100 scale, since improvement in OR models are precisely not dramatic. https://openaccess.thecvf.com/content/CVPR2021/papers/Wang_S...

There's a reason to not compare it with the YOLOv4 family? If I recall correctly, the advantage of YOLOv5 over YOLOv4 is still disputed and YOLOv4-tiny seems widely used.

I went to a technical high school in Mexico. Here's divided as 6 years of grade school, 3 of Middle school and 3 of High school.

After finishing this "bachelor" high school you get a technician degree. Mine was an electronics technician. And well, I met some of my best friends there and equals even when most of them went to different majors in college.

I think is great when you can get specialized subjects in such environment since people with similar interests can interact between them more easily.

Granted, there were bullied people, others that just wouldn't fit, but for the most of us (on classrooms over 40 students) we went along fine.

And certainly I felt it a better environment than middle school (12-15yo). Friends that really didn't like it, weren't that interested in electronics or science, but still we got along.

But well, in the end, Mexico is a place where we care about our relationships. Maybe too much.

I don't think the concept of high school is wrong. On any country contacts are important and usually change your life, and any social hub in life is a chance. It's just that, environment and motivation matters.

Graphics also matter. My 6yo Inspiron let me develop with 8GB of RAM, last year I bought a new one with Ryzen5 set the very same link setup and even used the same SSD but it would began to freeze or crash...

The APU was reserving 3GB of RAM for the GPU. The next day I went for 8GB more.

I think I did that, then have problems with the windows installer. A driver prompt. I regularly have to give services to laptops, so I prefer options that "just works". This is also the reason I don't usually use the Windows installation media creator, because is slow and won't burn my several ISOs. There was a time the new Win10 iso installation process would fail on any configuration I would try... I lost somewhat around THREE DAYS trying to solve each f*"** cryptic error message and an old iso did the trick. Just a few hours updating and it was done.

Sadly most of my clients won't just accept the option of installing Linux ;)

Windows images that are over 4.3GB(I don't remember the exact max size) most image burning tools will fallback to NTFS on the USB formatting and some motherboards/laptops don't like this. (Even on my rather modern Dell laptop)

The best solution so far I've found is to manually partition the USB into a 1GB FAT32 partition and the rest NTFS. Then you extract the files where needed [0]

I learned this method just last year but it seems it's pretty universal. At least for Win10 ISOs. And I don't have to boot Windows and wait for its updates!

[0] https://techbit.ca/2019/02/creating-a-bootable-windows-10-ue...

There are available eInk with a DSI interface and Linux drivers? It shouldn't be that hard to build a DIY eInk slim tablet with a PiZero or for some, with a CM4.

Same problem with latin american keyboards found just a few kilometers south the US. If I remember correctly, the default modifier key in VIM is next to imposible to press on a l-a keyboard. It was ` ?

One way to solve 3 and maybe 2 would be adding to the ecuation an ESP32/8266 and use it as an access point for the lamps. Then you might create any physical controls for the lamps or with some network magic add it to your infrastructure through a segmented network. I'm not sure if this can be done with an ESP alone (hence "network magic") or you could just use a second ESP connected to your private network and passthrough your commands via a serial port to the Lamp's ESP AP.

ESP32's are fairly cheap, easy to use and can even be programmed through micropython.

Yeah! That has been my idea for a long time! When thought about it, I began looking into devices that can output DC and charge a backup battery but with a variety of input power sources, Solar as main source and line AC as backup. Didn't get that many practical search results.

But it seems what I was looking for are boat chargers. They're meant to provide energy through solar/battery and charge at the same time while parked on a deck. Didn't followed it up though.

Especially, but not exclusively during unexpected power loss

Though in my experience as a technician installing 2.5" consumer grade / entry level Kingston/ADATA SSD to several clients, this is what generally will kill NAND flash devices, it's not an exclusive phenomenon to uSD cards.

I have a few running RPi's on premises for years. Ranging from v1 to now v4, certainly the only uSD 3 failures I have had has been with the early models and reliability went up with...... just a No-break power source. Just that.

It seems a lot of people think a sudden power loss it's what's only dangerous for electronics and NAND flash devices but for devices that are highly sensitive to browning down power rails (specially those first v1 raspis) any unstable power source is a rave party from which is hard to get out unscathed.

Once I found myself reading about device module programming since the common USB-Serial device module (I forgot its name, cdc something) wasn't properly working for a Chinese multiserial port chip inside a GSM cluster modem (one USB device to multiple serial ports).

I was attempting to hack away a simple example but I found the USB-Serial (more) generic driver intended for "test only" and... it just worked.

Another reason for reading about IO calls, schedulers, etc? That "I'm still writing data into your USB flash drive even when the GUI says it finished 5 minutes ago" that I hate so much.

What kind of SSD would be small? Those with a M.2 slot? Or there are actually small form factor SATA cable SSD's? Either way, you would need an extra USB adapter I guess.

EDIT: Regarding the heat sink, with the numbers some reviewers have shown shouldn't there be enough material to shed and still having it working fine overclocked? Granted, you lose the very well designed heat sink, but that's part of the hacking. It was not meant to work that way.

One thing that excites me the most is the hackability of the "included" case. Enough room for a battery, supercap, switches, normal HDMI ports, and still having free of clutter the GPIO pin rows.

I could also expect a 3D printed detachable LCD bezel/case connected directly to the microHDMI port and GPIO power rails. But well there are already some raspi based laptops for that matter.

I still like the R400 as a "presentation keyboard". Battery pack and HDMI adaptor and it's connected to the classroom projector.

by just pictures Intel spring pins are shorter and broader on the base side and with a steep rising angle. These look (quite) longer and with a lower rising angle. I guess these pins are more feeble and imprint less force than CPU spring pins.

I worry about the spring force of those pins after so many extreme temperature changes. Or AMD spring connectors design are similar to these?

EDIT: I meant "Intel spring pin connectors"

Actually as far as I know, photovoltaic cells are actually "common" diodes. And every diode actually emits light while forward biased, it's like a inherent property of them. Electrons "leap" the electron gap in a NP junction and while "going down" emit a photon. And the reverse is true, a photon of certain wavelength inciding on a NP junction will energize an electron if the wavelength coincides with the electron gap, thus producing a voltage difference.

This is used in more interesting ways on LEDs since their construction exposes their junction to the environment. You can use their small junction capacitance while they're charged and it's variability while being exposed to photons to create "light based touch sensors". You emit the same wavelength light with contiguous LEDs and one of them is "turned off" and time it's discharging rate. Faster? Light is shining over the junction. Lower? there's no matching photons over the junction.

So, I guess yes, is like PV cells but for even wider wavelengths (AFAIK PV cells already try to harvest the most of the sun's infrared light)

I had one similar realization but more focused on learning. "What's the difference between an expert and me?" Like a tennis pro, or a recognized programmer (besides muscle memory) "Information" Even emotions could be seen as a state of certain information in your head. Even your ego. "We are not that different. I can learn anything."

This doesn't count efficiency and optimization. It was just that given enough time and resources we are capable to "do something".

It's hard to learn that hard work will beat aptitude for a skill and that we actually have limited time in our lives though. But as you mention it makes a difference in the mind of people. The mental blocks many people have against learning new things really hinders how people interact with this rapid changing modern world and with their social groups. Also, there's these kind of realizations that humbles you (or lift your spirit up) and helps you being less egocentric, more empathic and forgiving.

I think bringing down these walls is a really hard problem.

There is work towards that end in Mexico. A selected few got a Supreme Court ruling that supports the legal use and growth of marijuana for personal use [1]. It didn't make it legal, but it sets precedence and a framework to keep the wheel moving to eventually make it legal like in other countries. Through this very same process last year the whole country make it same-sex marriage legal.

[1] https://www.nytimes.com/2015/11/05/world/americas/mexico-sup...

I wouldn't invest (not even my time) in licenses and staff to work over a platform that won't be maintained in the next 5-10 years. Stability is better for business.

When you have a massive leak, it's better to have under 10 angry users about that they have been hacked (increasing at a slow pace) than most of your userbase raging at you and leaving your service.

I guess they might be also worried about the low literacy on computer knowledge among their citizens.[1]

More mandatory programming courses could increase existence of desktop/laptop in households. The reference I included says that most people prefer smartphones than desktops/laptops, and programming could be a good enough reason to make them use them.

1. https://news.ycombinator.com/item?id=11758933

With wire-transfers you mean international wire-transfers? I think it might be just an american thing, mexican banks usually don't charge you anything for wire-transfers inside the country (in their "use your credit/debit card at least once each month plans") and well, it's Mexico. Though I think "first world"-something is quite the phrase just to describe consumer comfort: not advanced, not better (or worse), just "it's ok, I don't mind".

---EDIT---

Wire-transfers between national banks are free if you wait a day; there's a small fixed fee if you want them immediately. Transfers between different clients in the same bank are free and immediate. And it's becoming very common to be able to pay a quite assorted list of services through your e-bank account (from utilities to departmental store accounts) without fees.