v0.0.3
·
70 commits
to main
since this release
27c6de4
This commit was signed with the committer’s verified signature.
dogancanbakir
Dogan Can Bakir
Changelog
- 65e7f33 Add balanced class weights, body text features, confusion matrix
- 1431a2a Add crawl command to dit-collect
- 569ceca Add dit-collect tool for page training data collection
- 798b03d Add page type classifier with three-stage ML pipeline
- dee3849 Attach model.json to GitHub releases
- 0b02aad Auto-download model on first run from GitHub releases
- 0c931a7 Bump crate-ci/typos from 1.43.1 to 1.43.2 (#2)
- 834515a Collect initial page training data (61 pages, 10 types)
- 3624549 Crawl 500 pages from sites.txt (754 total)
- 3db7cd3 Download model from repo instead of GitHub releases
- 7338faf Expand page training data to 255 samples (10 types)
- 27c6de4 Fix golangci-lint errors in dit-collect
- 42a4309 Ignore dit-collect binary
- 7719c29 Optimize page evaluation: train form model once, remove body text pipeline
- df4b026 Retrain model with 754 page annotations
- 0754e56 Retrain with body text pipeline
- 0389e4c Track all data/ with git-lfs
- a407a53 Track page data with git-lfs
- c96aae9 Train page type model and move seeds into data/pages/
- a9deb43 Update README and CONTRIBUTING for page type classifier
- cda4c12 Update docs for page type classifier
- c7cf9f7 init