Claude
6bf250a086
Build: a draft far below the best keeps its share of the odds and the word constructor is no weight; Upscale with a model file waits for ComfyUI's queue
...
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com >
Claude-Session: https://claude.ai/code/session_01Tw9fjVsBg3ixjnT1Fgdy12
2026-11-08 16:39:59 +00:00
digital-garbage and Claude Opus 5.5
5f0b904530
feat: Build's like-guess learns which words you like; turns on only once it calls past ratings right
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com >
2026-10-10 00:46:48 +03:00
digital-garbage and Claude Opus 5.5
57c74268cd
fix: ratings set the odds (each like up, each dislike down, a run of dislikes much more); a redraw never commits mid-swap
...
A rating keeps the prompt that made the render and when; matcher replays the expander exactly.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com >
2026-10-10 00:13:50 +03:00
digital-garbage and Claude Opus 5.5
7917e4070e
fix: ratings keep teaching after a regenerate, a cleared text or a left-out clip; Build never assembles a rated-down shortcut or pair
...
Flush waits for a save already on its way; a field removed by a redraw no longer commits on blur.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com >
2026-10-09 23:46:19 +03:00
digital-garbage and Claude Opus 5.5
7b52643271
fix: Build weighs ratings only, matches triggers as the expander does, never wipes a scene; held arrows commit on release
...
Stats read 6000 scenes x 2400 triggers in under 1 s (was 2 min). Length fields show the grid value they use.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com >
2026-10-09 17:59:01 +03:00
digital-garbage and Claude Opus 5.5
e877ad08b7
rework: Build tab drafts a scene from your best-rated shortcuts; learns from rated scenes in projects
...
Drops the run log, edit-replay swap and Generate capture calls (core no longer depends on the feature).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com >
2026-10-09 17:18:15 +03:00
Claude