19 Commits

Author SHA1 Message Date
William Jeynes 77cdd9a01c Add statistics for model experiments. Fix dead link in documentation 2026-03-25 21:57:52 +00:00
William Jeynes 80bc151379 add majority voting 2026-03-24 16:50:41 +00:00
William Jeynes 5ce64290ce Make an ensemble model to combine scores together (very high accuracy) 2026-03-24 15:50:41 +00:00
William Jeynes 8c1e35f66f Increase dropout on regression model to cut down on overfitting 2026-03-24 13:16:18 +00:00
William Jeynes 44395bb251 add linear regression model initial version 2026-03-24 12:25:15 +00:00
William Jeynes e368c50577 Add training scripts for distilled, flan. Add run service for flan 2026-03-23 22:43:59 +00:00
William Jeynes c69730df6b Refine scoring to allow for better iteration on frontend. Update generate_adversarial.py 2026-03-22 16:04:38 +00:00
William Jeynes f4e84af272 Make the model less overfitting. Make it harder for an event to be classed as "perfect" 2026-03-18 01:05:24 +00:00
William Jeynes 8052d5c7ba Working on making the classifier harsher on unseen data 2026-03-17 22:19:03 +00:00
William Jeynes b08c1ada70 Small changes for the next set of human ranking 2026-03-17 00:18:32 +00:00
William Jeynes c89c7054fe Update agent to support new verification style. Update frontend to support new file format and remove redundant logic from old experiments. 2026-03-16 17:16:58 +00:00
William Jeynes 0a7bb114d2 Add removing of duplicates from pipeline. Add to sort step. Move score logic to robertaMetrics node. 2026-03-13 14:51:14 +00:00
William Jeynes d5c6cb444d Add better scoring, ignoring duplicates, catching under and over confidence. Showing difference between "FINE" and "PERFECT" 2026-03-13 12:18:52 +00:00
William Jeynes 8311556855 Add ROBERTA classifier ranking PoC, with 77pc off the bat 2026-03-13 11:24:51 +00:00
William Jeynes f09e36e740 Add initial version of ROBERTA classifier, add ability for multi pi charts 2026-03-11 22:02:31 +00:00
William Jeynes f14d112017 Add difference between auto scoring system and our own labels 2026-03-03 15:58:39 +00:00
William Jeynes 6ae551a93f Ensure date is passed to pipeline. Fix woring 2026-03-02 14:58:26 +00:00
William Jeynes c94812ed80 Prepare for mass data collection. Reduce concurrency as to not overwhelm scraper on long sessions. Remode duplicates from fetch script. Removing naming wierdness on scorer frontend. 2026-02-27 14:41:10 +00:00
William Jeynes 201176e71c Refactor scorer for future maintainabiliy 2026-02-26 10:25:49 +00:00