`models/LICENCE.md` recorded, on 2026-08-21, that no YOLO model trained on
ADE20K existed in usable form — the stuff classes photography cares about,
sky and vegetation and water, had no model to come from. Re-checked
2026-08-30: Ultralytics now ships a `semantic` task with ADE20K
checkpoints, so `models/scene/` holds `yolo26s-sem-ade20k`.
This is an addition, not a replacement. A semantic model labels every
pixel but merges same-class pixels into one region, so it cannot tell
three people apart — which is exactly what clicking a subject needs, and
exactly what `segment/`'s COCO instance model already does. The scene tab
grades per category and does not care that instances are merged. Keeping
both is the point.
## The export is truncated, deliberately
Ultralytics ends the graph with `Resize -> ArgMax -> Cast` and hands back
a `[1, 640, 640]` u8 label map. The script cuts that tail and exposes the
classifier's `[1, 150, 80, 80]` f32 logits instead, for two reasons.
Cost: the Resize materialises 150 x 640 x 640 x f32, 246 MB, and ArgMax
then reduces across the channel axis, striding 409,600 elements per
comparison. On one loaded machine the full graph ran ~1160 ms against
~500 ms truncated — roughly four fifths of the time spent on work the
application discards. Those numbers were measured under contention and
are upper bounds, but the ratio is structural.
Softness: ArgMax destroys the per-class scores, and the scene tab needs
them. Softmax over the 150 channels, summed within each photographic
category, yields per-category weights summing to 1 at every pixel.
Feathering a partition of unity cannot double-grade a boundary, whereas
feathering hard labels outward from two adjacent categories paints both
grades into the overlap and haloes every horizon.
The discarded upsample was never information: the graph's true spatial
resolution is the 80x80 logit grid, and the application can resample from
that itself.
The tail is matched by op type and asserted before cutting, so an
upstream graph change fails loudly in the exporter rather than quietly
shipping a differently-shaped model.
Nothing reads these weights yet — the decode path, the category
descriptor grouping 150 classes into ~8 photographic ones, and the scene
tab are still to come. At 24 MB this model also wants the runtime-asset
treatment `models/face/` already gets on Android rather than
`include_bytes!`; embedding it would put ~35 MB of weights in the binary.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
`mktemp -d` lands in `/tmp`, which on this and most current Linux
distributions is a tmpfs — memory, not disk, sized at half of RAM. The
venv this script builds installs torch into it, several gigabytes, and
the failure mode is not subtle:
error: Failed to install: torch-2.13.0-...whl
Caused by: No space left on device (os error 28)
on a machine with 102 GB free on the filesystem holding `/var/tmp`. The
quieter version of the same bug is worse: when it does fit, it evicts
whatever the user had in page cache to make room.
`${TMPDIR:-/var/tmp}` respects an explicit TMPDIR and otherwise picks the
disk-backed directory, which is what a multi-gigabyte throwaway wants.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The weights were in two places: face detection and recognition in
`models/face/`, segmentation in `core/dr-segment/models/`. Nothing was
wrong with either path, but between them there was nowhere to look to
answer "how much model does this application carry", and that number is
about to start growing.
So the crate-local copy moves up beside the other. `models/` now holds
`face/` and `segment/`, and a `du -sh` of one directory is the whole
answer.
No content changes: the .onnx and its vocabulary are byte-identical, and
`LICENCE.md` moves up a level to cover the tree rather than one crate.
The LFS pattern in `.gitattributes` is `*.onnx` and already matched both
locations, so only its comment needed the new path.
`include_bytes!` is relative to the source file and `build.rs` runs with
the crate root as its working directory, which is why the two paths climb
a different number of levels.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Comparing a larger model against a larger input size meant re-exporting at
resolutions other than the shipped 640, and the script only ever wrote that
one number. `IMGSZ` is now a second positional argument, defaulted to 640 so
every existing call is unchanged.
The experiment this was built for found bigger input a net loss on its own
merits — yolo26n-seg and yolo26s-seg at 1280 both lost track of large,
frame-filling subjects (a bus's box shrank and its score nearly halved)
in exchange for catching small or partially-occluded ones tiling already
handles. Nothing shipped from it, but the ability to re-run that comparison
is worth keeping.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Local masking needs to know where an image's regions are. The watershed
spike (S15 arm A) found the boundaries but had no idea what any of them
enclosed; its coarse levels were geometric accidents. This adds the other
half and the thing that joins them.
`core/dr-segment` is where region reasoning now lives — the hierarchy moves
out of `dr-gpu`, which keeps only the pixel passes that are genuinely
shaders. The new crate is device-free and, without its default features,
model-free too: 20 of its tests need neither an adapter nor 11 MB of
weights.
Arm B runs YOLO26n-seg through `ort`. D13 framed inference as a choice
between `ort`'s C++ runtime and the pure-Rust dependency policy; that was a
false choice. `ort`'s `alternative-backend` feature unlinks the C entirely
and `ort-tract` supplies the API from tract, which is pure Rust. Measured
before committing to it: zero unsupported operators, 420 ms for 640x640,
and correct masks on bus.jpg. No NDK problem to solve, so D13's largest
tolerated exception is not needed.
Arm C is `prior.rs`, and it ships because the two arms fail in opposite
directions. Instance membership re-weights the merge saddles, so region
pairs the model believes share an object merge early and pairs straddling
its edge merge late. No boundary moves — only the order in which they
dissolve — which is how the result stays pixel-accurate at every level
while its coarse levels become named things.
Two things the spec assumed that turned out to be false, both recorded in
models/LICENCE.md: there is no usable ADE20K-trained YOLO, so the shipped
vocabulary is COCO's 80 subjects and *stuff* like sky and foliage must come
from arm A; and tract cannot parse a dynamic-shape export, so the graph's
input is fixed and tiling is the only route to more semantic resolution.
Weights are AGPL-3.0, which GPLv3 §13 permits and which makes the combined
work effectively AGPL. Deliberate, not accidental. They live in Git LFS,
and a build script fails with an instruction rather than embedding a
pointer file when the clone lacks them.
dtourolleandClaude Opus 5