Image-Tagger is a simple command line tool to tag images using exiftool. It uses AI to detect the content of the image and add tags to the image metadata.
AI engine used is Ollama and the vision model is LLava and the llm model is Phi3.
Install Ollama and the required LLava, Phi3 AI models.
ollama pull llava
ollama pull phi3
To install Image-Tagger, simply run the following command:
pip install --user .
or with an isolated environment:
python -m venv ~/bin/image-tagger_venv
~/bin/image-tagger_venv/bin/pip install .
ln -s ~/bin/image-tagger_venv/bin/image-tagger ~/bin/image-tagger
To use Image-Tagger, simply run the following command :
image-tagger --help
To contribute to Image-Tagger, simply fork the repository and create a pull request. Please make sure to include a detailed description of your changes. Here are the things I will check during the review :
- Is CHANGELOG.md have been updated (required)
- Is the lint score did not decrease (required)
- Is the test coverage did not decrease (required)
- Is the documentation have been updated (if required)
- If tests have been added (optional)
Pour définir le nombre maximal de modèles chargés simultanément à deux, utilisez :
launchctl setenv OLLAMA_MAX_LOADED_MODELS 2