• 0 Posts
  • 26 Comments
Joined 5 months ago
cake
Cake day: June 4th, 2024

help-circle





  • Darohan@lemmy.ziptoLinux@lemmy.mlLinux middle ground?
    link
    fedilink
    arrow-up
    11
    arrow-down
    1
    ·
    2 months ago

    This may be an unpopular opinion, but NixOS. It has package up-to-dateness comparable to (and sometimes better than) Arch, but between being declarative (and reproducible) and allowing rollbacks, it’s much harder to break. The cost is, of course, having to learn how to use NixOS, as it’s a fair bit different to using a “normal” Linux distro.





  • Congrats on expressing that in the most passive-agressive and gatekeepery way you could’ve. I’ve been using Linux for the better part of a decade now, and know my way around the usr dir - however things work a bit different on NixOS, whose package manager doesn’t involve installation steps beyond adding the word “helix” to my packages list. I’m not great at reading though, so I absolutely would’ve missed something as obvious as the Installation page 😅 As for your beliefs about postmodern Vim clones, what’s the point (and fun) in the freedom of choice Linux offers if I can’t install and try out the latest fun spin on an old fave from time to time?




  • Was ready to downvote but this is actually a really good guide, well done OP! The one issue I will raise, though, because I faced it myself, is that as long as you’re still using Windows, it is way too easy to just go back to using the Windows programs not the open source ones. Only through switching to Linux can you really “throw yourself into the deep end” and force yourself to learn these new things. Microsoft has made themselves the “path of least resistance” (or at least that of “most momentum” for a reason) and if you’ve been using a computer for a while, it’s a lot easier to break the habits and realise the benefits by giving yourself no other option than it is by trying to discipline yourself into using the new options.




  • Just gone through this whole process myself. My god does it suck. Another thing you’ll want to be aware of around Takeout with Google Photos is that the photo metadata isn’t attached as EXIF like with a normal service, but rather it’s given as an accompanying JSON file for each image file. I’m using Memories for Nextcloud, and it has a tool that can restore the EXIF metadata using those files, but it’s not exact and now I have about 1.5k images tagged as being from this year when they’re really from 2018 or before. I’m looking at writing my own tool to restore some of this metadata but it’s going to be a right pain in the ass.


  • Darohan@lemmy.ziptoScience Memes@mander.xyzSardonic Grin
    link
    fedilink
    English
    arrow-up
    1
    ·
    4 months ago

    I feel like you’re viewing this from the wrong angle, or at very least we’re viewing it from different angles. You seem to be doing a binary classification (Is this plant edible) rather than a group classification (what plant is this?) where edibility is an attribute of the plant to be returned to the user (yes; no; when green; only the roots; etc.) - the latter is the approach most of these apps take, classify the image into a species (or list of potential species) then give the user details such as identifying features, common growing areas, edibility, and lookalikes. You’re right about softmax, it’s been a couple of years since I’ve done the programming side of this so my terminology is a bit rusty.


  • Darohan@lemmy.ziptoScience Memes@mander.xyzSardonic Grin
    link
    fedilink
    English
    arrow-up
    10
    arrow-down
    1
    ·
    4 months ago

    This is blatantly false. Classification tasks like this all have a level of certainty for each possible category - it’s just up to the person writing the software to interpret those levels of certainty in a way that’s useful to the user. Whether this is saying “I don’t know” when the certainties are too spread out, or providing a list of options like other people in this thread have said their apps do. The problem is that “100% certainty” comes off well with the general public, so there’s a financial incentive to make the system seem more certain than it is by using a layer (from memory it’s called Softmax?) that will return only the category with the highest degree of certainty.



  • If I had to guess that’s gonna be a quirk of ActivityPub, and should self-resolve in a little bit, but I’m not an expert so don’t take me at my word there. I have some experience self-hosting setting up my own homelab over the last 2-3 years - if you’d like some “getting started” conversation, feel free to send me a DM or contact me on Matrix @darohan:tchncs.de