config.json: v1.3.0 license-screened retrain numbers

License-screened retrain (v1.3.0): new head trained on the 3,298-document corpus; card numbers updated

Sync top-feature coefficients to the released fit

Stacked two-stage head (held-out 92.9%)

Config for 1.2.0

Stacked head release 1.2.0

Document the selectable PDF backend (pdfplumber default, pymupdf opt-in)

Update corpus deployment rate to the pdfplumber pipeline (76.4%)

Release 1.1.0: MIT-clean pipeline (pdfplumber + pypdfium2)

Release 1.0.0: head trained on the human-validated corpus

Add AI-generated disclosure to image hover text

Add cross-references to GitHub repos and training dataset

Soften training data verbiage: posters are repository-labeled, not verified

Add config.json for HuggingFace download tracking

Updated README with logo, pipeline diagram, related models, grant acknowledgment

Add PosterSentry logo

Upload pyproject.toml with huggingface_hub

Upload scripts/train_poster_sentry.py with huggingface_hub

Upload src/poster_sentry/classifier.py with huggingface_hub

Upload src/poster_sentry/__init__.py with huggingface_hub

Upload src/poster_sentry/features.py with huggingface_hub

Upload models/poster_sentry_head.npz with huggingface_hub

Upload README.md with huggingface_hub

initial commit