Autotagger problem

Started by jln, August 15, 2025, 06:08:37 PM

Previous topic - Next topic

jln

I was able to use autotagger without difficulty in the previous version, but I am having trouble with the new version. I have carefully read the help on autotagger, but apparently missed what I needed to know.

I thought I would try by seeing if I could get anything out of Ollama. When I have tried running it, I get a message that there is a problem that is shown in a category. The category panel does show a problem, but doesn't explain what it is.

I am attaching the log file. Let me know what I have done wrong. Thanks.

Mario

It seems you have not installed the model in Ollama. Ollama reports

{"error":{"message":"model \"llava\" not found, try pulling it first"

Did you follow the instructions here: Ollama(Local AI)?
There is also a video tutorial in the IMatch Learning Center: https://www.photools.com/imatch-learning-center/ (scroll to the end).

The LlaVA-Model is older, I would recommend to install the Gemma 3 model as explained in the help topic.
If your graphic card has 8GB VRAM or less, use the 4b Variant, else the 12b Variant for better results.

On the command line, run

ollama run gemma3:4b

or

ollama run gemma3:12b

to install the Gemma model in Ollama. Then you can use it from IMatch AutoTagger.