p.s.: For those who are (understandably!) sceptical about AI, here is an excerpt from the FAQ:

But isn't using AI bad for the environment?

[…] not everything utilizing AI has to do with LLMs running on […] server farms owned by billionaires. STAG uses a very small convolutional neural network (CNN) which does not even need a GPU to run fast and efficiently. You can run STAG on a perfectly normal computer and not draw more power than your Adblocker needs for making the internet bearable.

Interesting thing. Hopefully I'll find the time to try this. I think DigiKam can use XMP sidecar files, too. (I use DigiKam.)

I always loved tagging in my former professional life. 🤪 (Name a computational linguist who doesn't love tagging.)

I'm going to give it a try, from what I saw from the recognize-anything repo it can roughly identify birds (Well at least if it's a heron, a falcon...) so it might be useful to sort my bird collection

I'm having some difficulty using this. Is there somewhere I can provide details etc?

Possible to get OSX 86_64 executable?

Ok, will give it a try

Okay, I get it now. You *did* note the one time download of the model file. I just didn't recognize it as such. My suggestion would be a note right after the message "Starting Tagger", you add a line along the lines of "Downloading Model File.

You know, for the slightly dim like me! ;) I'll add a note to y Github Message as well.

Maybe you could also add the option to specify the directory where the recognize-anything model is stored. I would like to have those 3 GB at some other location than the current .cache directory on Windows.

oh wow. I've been considering, and putting off, tagging my 10000 or so photos. This looks like it will be extremely useful. Thank you 😊

ERROR: Could not find a version that satisfies the requirement torch==2.6.0+cpu (from versions: none)
ERROR: No matching distribution found for torch==2.6.0+cpu

$python3 -m pip install torch torchvision torchaudio

ERROR: Could not find a version that satisfies the requirement torch (from versions: none)
ERROR: No matching distribution found for torch

Any ideas how to proceed? Web searching says it should just work… (2/2)

Always seems to fail at installing pip code when in venv. torch won't install I know this is not an issue with your code, but I've tried many other executables with the same issue. Running python 3.13.1
$pip install -r requirements.txt

Collecting tkinter-tooltip==3.1.2 (from -r requirements.txt (line 10))
Using cached tkinter_tooltip-3.1.2-py3-none-any.whl.metadata (8.0 kB) (1/2)

This looks really useful. Thanks for sharing it.

Hi

Same error
Collecting rawpy>=0.23.1 (from -r requirements.txt (line 8))
Using cached rawpy-0.24.0-cp313-cp313-macosx_10_9_x86_64.whl.metadata (6.2 kB)
Collecting pyinstaller==6.12.0 (from -r requirements.txt (line 9))
Using cached pyinstaller-6.12.0-py3-none-macosx_10_13_universal2.whl.metadata (8.3 kB)
ERROR: Could not find a version that satisfies the requirement torch==2.6.0+cpu (from versions: none) (1/2)

interesting! Peakto offers AI tagging. I tried it several times but the quality was just laughable bad.

Interesting!

different, but same

early in install:

a13e38cca6655f83ad0185271167dbcbf
Installing build dependencies ... done
Getting requirements to build wheel ... done
Preparing metadata (pyproject.toml) ... done
Ignoring torch: markers 'sys_platform == "linux"' don't match your environment
Ignoring torchaudio: markers 'sys_platform == "linux"' don't match your environment
Ignoring torchvision: markers 'sys_platform == "linux"' don't match your environment… (1/3)