Upstream:
- 🚀 LibrePhotos: You can now download a large language model to post process captions
Now you can add previous attained knowledge to captions like names and places. This is a setting, which should only be used be high performance systems with at least 16GB RAM, because large language models are cutting edge tech.
- 🚀 LibrePhotos: You can now see the image tag on the sidebar
- 🚀 LibrePhotos: Tags in a caption will now add a new album in things
- ✨ Frontend: Generated captions will now be shown as a suggestion
- ✨ Frontend: Allow setting heavyweight worker higher than 3
- ✨ Frontend: Added an edit mode for captions
- ✨ Frontend: Face Labeling List is improved as entries render as buttons with a hover effect
- ✨ Frontend: Added a warning when selecting a large captioning model
- ✨ Frontend: Already existing tags will now get suggested, if you enter #
- ✨ Frontend: Implemented a rich text editor for nice rendering of tags in caption textarea
- ✨ Frontend: Continuation of migrating to RTQ (Implemented by sickelap)
- ✨ LibrePhotos: Allow searching for lenses like "front camera"
- ✨ LibrePhotos: Calculating Clip Embeddings is now a separate service
- ✨ LibrePhotos: Image captioning is now a separate service and unloads after 30 seconds
- ✨ LibrePhotos: Updated around a bunch of dependencies and updated language strings from the community
- 🔨 Frontend: Changing the person in the lightbox, will now refetch correctly the new person
- 🔨 LibrePhotos: Do not change capitalization when saving a caption
- 🔨 LibrePhotos: Fix multiple issues when importing faces from Apple Smartphones
- 🔨 LibrePhotos: Download models when changing captioning model
If you are interested on how I develop this application, you can watch me develop features live on my channel
Sponsoring development:
My work is sponsored by 9 awesome people!
If you like the work I do, then you can support me via GitHub sponsors or via PayPal