UNesT — Worker pipeline v3 (batched-fast)

Whole-brain 133-class segmentation with batched-fast tiling: N tile slots alive simultaneously, each forward scatters to all overlapping tiles in one pass. Worker auto-picks N based on the GPU budget (M1 16GB UMA → N=8, ~210 s vs v2's 607 s — Node-verified at byte-exact 99.9999%).  ·  ← v2 (outer-loop tiling)  ·  v1 (main-thread)
✓ MONAI strict conformance + 2.89× speedup: whole-volume argmax 99.9999% vs MONAI SlidingWindowInferer (Node verified, MPS reference). Pipeline: NormalizeIntensityd(nonzero=True) → (Z,Y,X)↔(X,Y,Z) transpose → batched-fast tile sliding (96³, overlap 0.7, constant). Worker auto-verifies against unest_full_ref_argmax.bin if present.
⚠ Download size: ~340 MB of UNesT weights will be fetched below.  Memory: N=8 batched-fast peaks at ~12 GB GPU buffers — needs Apple Silicon 16 GB+ UMA (M1/M2/M3) or discrete GPU 16 GB+. Other platforms auto-downgrade to N=2/4 (or N=1 = v2 outer-loop fallback).

Log

Slice viewer

Normalized T1w
133-class seg
Overlay