https://github.com/dkackman/captioner
https://github.com/dkackman/captioner
Last synced: over 1 year ago
JSON representation
- Host: GitHub
- URL: https://github.com/dkackman/captioner
- Owner: dkackman
- Created: 2023-03-16T00:43:36.000Z (over 3 years ago)
- Default Branch: master
- Last Pushed: 2023-07-22T12:28:39.000Z (almost 3 years ago)
- Last Synced: 2025-03-14T20:45:34.274Z (over 1 year ago)
- Language: Shell
- Size: 22.5 KB
- Stars: 1
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
Awesome Lists containing this project
README
# Introduction
An experiment in using AI to manage and manipulate family photos.
Goals
- [x] Generically tag an entire directory strucutre of photos (a man and woman pose in front of a chirstmas tree)
- [ ] Train a model to recognize family members and pets
- [ ] Specifically caption an entire directory strucutre of photos (Don and Daisy pose in front of a chirstmas tree)
- [ ] Make an image search and organizer tool on top of that
- [ ] Create a stable diffusion LoRA from the specifically tagged photos (imagine a picutre of Don and Daisy in front of a christamas tree on USS Yorktown)
## Install
Make sure you have left the venv.
```bash
sh install.sh
. ./activate
```
## Scripts
### `01.prepare.py`
Copies image files from a known location, converting them all to jpg and filtering our thumbnails (by file size) and corrupt files. It also flatten any directroy structure and just gives every file a uuid name. (optional)
### `02.label.py`
Runs image-to-text cpationing on an directroy of photos. Stores the generated caption in a csv file.