Oct 17 Today 18:25 Remove weights/model.yaml - superseded by the single root config.yaml 18:25 Load the model config from the single root config.yaml instead of weights/model.yaml 18:25 Trim dead training-only code paths (loss-alpha bookkeeping, unused random-init branch, dead pre_load_model checkpoint-surgery block, unused imports) - none of it executes for inference; verified via strict safetensors load + identical forward-pass output 18:16 Replace AV-WIDE example with a better-lit source image (the old one was nearly black, vessels looked like they were floating on nothing) 18:04 Add root-level config.yaml so HF Hub counts downloads (its default download-tracker only watches config.json/config.yaml/etc. at repo root) 09:26 Link the RLAD + PVBM end-to-end tutorial 09:20 Note bfloat16 inference precision in Quickstart 09:20 Load and run inference in bfloat16 (the precision the model was trained in) 07:42 Fix: model output is already at content size (1444), crop_padding was incorrectly double-cropping it to 1416 and misaligning the mask 07:39 Rehost PVBM optic-disc-segmenter weights (Google Drive link was broken/rate-limited) Oct 16 Yesterday 22:16 Link both RLAD and REYIA in the intro sentence 22:16 Reorder: intro, qualitative examples, quickstart, example, results, ablations, license, citation 22:10 Add explanatory sentence before Table 1 22:09 Condense intro and qualitative-figure section to one sentence each 22:07 Move qualitative figure right after intro, remove How it works section 22:07 Regenerate example with higher mask alpha 22:07 Add mBRSET, reorder with AV-WIDE last, higher mask alpha, native aspect ratios throughout 22:04 Fix overlay to preserve native aspect ratio instead of stretching to square; raise mask alpha 22:01 AV-WIDE: show at native aspect ratio instead of stretched to square 22:00 LES2 is ENRICH; link UZLF to KU Leuven RDR, others to REYIA's own license detail 21:59 Add Tables 3-4, stop asserting UZLF's release status without evidence 21:59 Add Table 4 (pretraining/generation ablation) 21:58 Add Table 3 (dataset ablation) 21:54 Punch up the intro 21:53 Add benchmark example predictions section 21:53 Add benchmark dataset example predictions 21:53 Lead with paper title, add results table, concrete license statuses, trim to match REYIA's voice 21:48 Fix naming: RLAD + SwinV2-Large is primary, LUNetV2 is the internal alias 21:48 Initial release: LUNetv2 SwinV2-Large retinal AV segmentation, RLAD-augmented 21:47 initial commit 
