Upload gc_submit/model_alldata_README.md with huggingface_hub

Upload gc_submit/model_alldata.tar.gz with huggingface_hub

GC submission backup: gc_submit/model.tar.gz

GC submission backup: gc_submit/reg2026_algorithm_image.tar.gz

Fixed container: remove h5py import + thumbnail dedup (patch_wsi_multiscale). Verified 20 raw WSIs = MetricA 0.8643 in rootless Docker.

Upload model.tar.gz with huggingface_hub

Upload reg2026_algorithm.tar.gz with huggingface_hub

Upload reg2026_daemonless_image_README.md with huggingface_hub

Add Claude Code session bundle (transcript+memory, token scrubbed) + RESTORE.md

Add method writeup (browsable HTML)

Add eval scripts (official scorer) + docs (method HTML + challenge pages)

Add 20 real test slides + features + container predictions (Metric A 0.8643 validation set)

Add deployable SOTA container (None-MIL + edge_disc, verified Metric A 0.8643 on 20 real test slides)

Upload reg2026_metricA_sota_README.md with huggingface_hub

Upload features_test/capped_1024_uni2_test.h5 with huggingface_hub

Upload features_test/capped_1024_10x_test.h5 with huggingface_hub

Upload features_test/capped_1024_test.h5 with huggingface_hub

Upload reg2026_exp_notes.md with huggingface_hub

Upload reg2026_metricA_sota_README.md with huggingface_hub

Upload reg2026_metricA_sota.tar.gz with huggingface_hub

cap=1024 MIL bundle: add 1020 multi-scale models (now 3 scales, 30 weights) + 10-slide scores in README

cap=1024 MIL models (TransMIL+ABMIL, 20x & 10x) + scripts + 10 test WSI

cap=1024 MIL bundle: 10 held-out test slide list

Fixed container: organ classifier + tiled memory-bounded patching + 5-TransMIL ensemble

Add model-card frontmatter

Model card: usage + model + training (organ-fix, 5-TransMIL ensemble)

REG2026 interf1 submission container (code+UNI2 weights+MIL+routing+sample WSI)

initial commit