√𝛂𝛋𝛆

√𝛂𝛋𝛆@piefed.world · 1 month ago

Something like, An image consisting of two men and three children, formally dressed, indoors, they appear to be at an event such as a wedding reception. There is a cake in the background and people dancing. There is a banner in the image that says "Congratulations Tom and Jerry".

√𝛂𝛋𝛆@piefed.world · 1 month ago

Scripting and the command line is an option. It may seem a bit daunting at first but can be a lot faster. Doing AI training stuff I tried brute forcing for awhile but it is just too tedious. A model like a Qwen 3 can be instructed to act like a photo captioning model. It is quite good at optical text recognition in images too.

You can script it to caption the whole image or any number of elements like color, location, subjects, etc., while also limiting the length of text. Then modify the image metadata with the text. It is also possible to feed it a set of json keys and it will fill in the values for the dictionary.

If you go this route, then you start using tools like ripgrep and it becomes possible to find and manipulate many thousands of images quickly to find niche sets in the tens to hundreds in just a single scripted command. You can do stuff like grep -ril foo | xargs | sed ‘s/foo/bar/’ or something like that. Not at my comp right now to pull up the command flags, but that is how to search and edit to change all instances of words in a bunch of files. That one is technically incorrect but it makes far more intuitive sense than alternatives.

Something like Qwen requires enthusiast level hardware, or if you have to, a free google colab instance. If you search for captioning models there are much smaller ones like the ancient BLIP that are less accurate but fast and run on anything. Anyways, I know this reply is not exactly what you wanted, but it is an option, and one I thought was beyond me until I tried it. Hopefully it is useful digital neighbor.

√𝛂𝛋𝛆@piefed.world · 3 months ago

They are nice for keeping tools around on spare SD cards that you might not want to run normally. Like that is a good way to look at Parrot or Kali Linux setups.

Checking out how to build an OS from scratch is also handy. It can be an interesting low risk way to explore building Gentoo, Arch, or Linux From Scratch.

The main appeal IMO, is that you have microcontroller like input, output, and serial communications already setup in the kernel with access in user space. As long as the kernel is supported by the Rπ foundation, (it is proprietary undocumented hardware that only they can support), you are getting the security updates required to keep the thing online automatically and safely. The best stuff to build is unique stuff for you that uses these aspects. Like make a little bathroom clock with a little TFT LCD display that tells you the local weather. Then set up some RSS feeds for local community stuff you do not want on your main mobile device, like maybe local political activity, library and community center events, concerts, clubs, etc.

For server stuff, I would stick with devices with purpose built hardware. Like, a micro SD card is slow and unreliable, and the lack of nvme is bad. In most cases it is cheaper to use other old devices that already have screens unless you want to share a hardware design that is repeatable, you need something secure to keep online, or you need serial or input/output. Those are the main benefits.

The thing is, the Rπ is what it is. It is the path of least resistance. The software support is approachable and great. The price is cheap. However, the non profit thing is a scam. The Rπ foundation is basically an arm of Broadcom. The Rπ is a chip from a set top TV box with 3/4 of the die unused. Broadcom uses excess fab capacity to make the Rπ chips and sell them at materials cost. This is not charity. It is controlling the grass roots market to make competitive scaling business ventures difficult. This is why Rockchip is not crushing them already. The Rockchip RK3588 chip is fully documented and open source. In this space, there is little to no innovation, it is only about price on ancient trailing fab nodes. This is the ladder to climb that leads to Intel, AMD, Samsung, and Qualcomm. The Rπ is the guy kicking anyone that tries to climb. So… use it for what it is good for, but in most cases, other hardware is better, and there is nothing wrong with saying so, or moving past the Rπ. I’m lying next to a RK3566 machine right now, sorting out issues with the ARM version of Fedora Workstation, looking at how to build the native WiFi module for it from source, and maybe try debugging an issue in that module’s code that is causing a memory race condition. Although that last one is past my typical pay grade. - So I’m not all fluff here.

That is just my $0.02.