Ship the three eye-state models beside the face pair

2d106det for the eye contours, OCEC for open or closed, SGC for
sunglasses — all three pinned to a batch of one by the same script as
the pair, and installed by every packager so the eyes-open filter works
out of the box. The two classifiers are MIT, code and weights; the
README records their provenance, SGC's undocumented training set, and
the hashes as fetched and as shipped.

apps/darkroom-android/src/lib.rs

@@ -324,11 +324,19 @@ fn unpack_bundled_models(app: &slint::android::AndroidApp) {
     // (`FaceDetector`, docs/faces.md §12.3) and a tablet has no other way to
     // obtain the one it was not shipped with. Twenty megabytes of APK for
     // the choice; the embedder is the same for all three.
-    const BUNDLED: [(&std::ffi::CStr, &str); 7] = [
+    //
+    // Then the three eye-state models (docs/faces.md §17): landmarks, open
+    // or closed, sunglasses. The app indexes without them; with them the
+    // eyes-open filter has something to read, and a tablet has no other way
+    // to get them either.
+    const BUNDLED: [(&std::ffi::CStr, &str); 10] = [
         (c"models/scrfd_500m_640.onnx", "scrfd_500m_640.onnx"),
         (c"models/scrfd_2.5g_640.onnx", "scrfd_2.5g_640.onnx"),
         (c"models/scrfd_10g_640.onnx", "scrfd_10g_640.onnx"),
         (c"models/arcface_mbf_b1.onnx", "arcface_mbf_b1.onnx"),
+        (c"models/2d106det_b1.onnx", "2d106det_b1.onnx"),
+        (c"models/ocec_s_b1.onnx", "ocec_s_b1.onnx"),
+        (c"models/sgc_l_48_b1.onnx", "sgc_l_48_b1.onnx"),
         (c"models/yolo26s-sem-ade20k.onnx", "yolo26s-sem-ade20k.onnx"),
         (
             c"models/yolo26s-sem-ade20k.classes.json",
@@ -343,8 +351,8 @@ fn unpack_bundled_models(app: &slint::android::AndroidApp) {
 
     for (asset_path, name) in BUNDLED {
         let dest = dir.join(name);
-        // Already unpacked. Not re-read on every launch: this is 61 MB of
+        // Already unpacked. Not re-read on every launch: this is 73 MB of
-        // copying across the seven entries, and the file does not change without
+        // copying across the ten entries, and the file does not change without
         // the APK changing, at which point the install wiped it anyway. It
         // matters more now than it did — a launch that skips every entry here
         // costs nothing at all, which is what makes the second launch after an

docker/windows/README.md

@@ -41,7 +41,7 @@ Measured on the first build, 2026-09-12:
 | DLL imports | 26 Windows system DLLs; **no** `libwinpthread-1.dll`, `libgcc_s`, `libstdc++` |
 | `wine darkroom-desktop.exe --version` | `darkroom-desktop 0.12.0`, exit 0, 0.1 s |
 | `makensis` | 105 MB `DarkRoom-<version>-x86_64-setup.exe`, 64-bit stub |
-| `wine setup.exe /S` | Installs exe + 7 models to `AppData\Local\Programs\DarkRoom`, writes the `HKCU` uninstall key; the installed exe runs |
+| `wine setup.exe /S` | Installs exe + 10 models to `AppData\Local\Programs\DarkRoom`, writes the `HKCU` uninstall key; the installed exe runs |
 | `wine uninstall.exe /S` | Removes the directory and the key |
 | Start Menu shortcut | **Not verifiable here.** `CreateShortcut` is `IShellLink` and does nothing under a headless Wine; the directory beside it is created. Check on Windows. |
 

docs/windows.md

@@ -281,13 +281,14 @@ $LOCALAPPDATA\Programs\DarkRoom\
   darkroom.exe
   models\
     scrfd_500m_640.onnx   scrfd_2.5g_640.onnx   scrfd_10g_640.onnx   arcface_mbf_b1.onnx
+    2d106det_b1.onnx   ocec_s_b1.onnx   sgc_l_48_b1.onnx
     yolo26s-sem-ade20k.onnx   yolo26s-sem-ade20k.classes.json   categories.txt
   LICENSE
   uninstall.exe
 ```
 
 Plus a Start Menu shortcut, and nothing on the desktop unless the user ticks it. The models are
-the same seven files the APK bundles and the PKGBUILD installs; `models\` beside the executable is
+the same ten files the APK bundles and the PKGBUILD installs; `models\` beside the executable is
 where §3.2's lookup finds them. **No `LICENSE` yet**: the repository has no licence file at its
 root (the Arch package points at the system's shared GPL text), so the installer has no licence
 page until one is added — a one-file change, and the `.nsi` says where the page then goes. The face weights carry the research-only grant that
@@ -458,7 +459,7 @@ container was the cheaper way to get a pinned MinGW and a Wine that could be thr
 | It links | Yes, first attempt once the link flags were right. 115 MB, `PE32+ … (GUI)`. Two dead-code warnings, both `cfg`-shadowed constants, fixed. |
 | It is a Windows executable | 26 imports, all Windows system DLLs. No MinGW runtime. `.rsrc` carries `PRODUCTVERSION 0,12,0,0`, `ProductName DarkRoom`, the icon. |
 | It starts | `wine darkroom-desktop.exe --version` → `darkroom-desktop 0.12.0`, exit 0, 0.1 s. |
-| The installer runs | `makensis` → 105 MB. `/S` installs the exe and seven models to `AppData\Local\Programs\DarkRoom`, writes the `HKCU` uninstall key; the installed exe runs; `uninstall.exe /S` removes directory and key. Shortcut unverifiable (§6). |
+| The installer runs | `makensis` → 105 MB. `/S` installs the exe and ten models to `AppData\Local\Programs\DarkRoom`, writes the `HKCU` uninstall key; the installed exe runs; `uninstall.exe /S` removes directory and key. Shortcut unverifiable (§6). |
 | It draws a window | Not attempted. |
 
 **What the first draft got wrong**, kept in place above with a note rather than rewritten, because

models/face/README.md

@@ -15,10 +15,19 @@ by every platform:
     scrfd_2.5g_640.onnx    3.3 MB   "Balanced" — 14% more faces for 12% more time (faces.md §12.3)
     scrfd_10g_640.onnx      17 MB   "Thorough" — a further 12% for 3× the time
     arcface_mbf_b1.onnx     13 MB
+    2d106det_b1.onnx       4.8 MB   106 landmarks, for the eye boxes (faces.md §17)
+    ocec_s_b1.onnx         483 KB   eyes open or closed, per eye
+    sgc_l_48_b1.onnx       6.1 MB   sunglasses or not, per head
 
 Which detector runs is a per-device setting (`FaceSettings::detector`); all three are installed so
 the choice exists on every platform. Each is its own `faces.model_id`, so changing it re-indexes.
 
+The last three are optional to the app: `library::face_models` reports them beside the pair when
+all three are there, and a library without them indexes faces and simply has no eye readings —
+every reader treats "never read" as unknown, never as closed. They are found in the *same*
+directory as the pair, so a hand-placed pair does not pick up a package's eye models from a
+directory it otherwise outranks.
+
 **A clone without git-lfs gets a ~130-byte pointer where each model should be.** Both packagers check
 for exactly that and refuse, rather than shipping the pointer and failing inside tract on the user's
 machine. Fix it with `git lfs pull`.
@@ -31,7 +40,50 @@ cannot parse any of the graphs while they are dynamic:
     ./tools/fix-face-model-shapes.sh det_2.5g.onnx  models/face/scrfd_2.5g_640.onnx --input input.1=1,3,640,640
     ./tools/fix-face-model-shapes.sh det_10g.onnx   models/face/scrfd_10g_640.onnx  --input input.1=1,3,640,640
     ./tools/fix-face-model-shapes.sh w600k_mbf.onnx models/face/arcface_mbf_b1.onnx --dim None=1
+    ./tools/fix-face-model-shapes.sh 2d106det.onnx  models/face/2d106det_b1.onnx    --dim None=1
+
+`2d106det_b1.onnx` is from the same `buffalo_l.zip` as `det_10g.onnx` and under the same grant: the
+106-point landmark model whose lid contours the eye boxes are cut from (docs/faces.md §17.2).
+sha256 as fetched `f001b856…a7109dbf`, as shipped `afc2984c…03368ef26`.
 
 The weights carry a non-commercial research-only grant. They are here because this is a private
 repository and self-installed builds; they come back out before anything is published, and the
 restriction binds whoever uses the app, not only the project. docs/faces.md §2.2a is the decision.
+
+## The two classifiers are a different matter
+
+`ocec_s_b1.onnx` and `sgc_l_48_b1.onnx` are **MIT, code and weights**, from Katsuya Hyodo's
+ultra-lightweight classifier series — the same author as the whole-body detector the reference
+pipeline uses. They are not under the InsightFace grant and do not come out when the project
+publishes. Read 2026-09-19:
+
+| | OCEC — open/closed eyes | SGC — sunglasses |
+|---|---|---|
+| Source | `github.com/PINTO0309/OCEC`, release `onnx`, `ocec_s.onnx` | `github.com/PINTO0309/SGC`, release `onnx`, `sgc_is_l_48x48.onnx` |
+| Licence | MIT (repository `LICENSE`; no separate grant on the weights) | MIT, likewise |
+| Training data | *Open and Closed Eyes* (Młodawski 2024, HF, **ODC-By 1.0** — attribution only), crops cut by DEIMv2-Wholebody34 (Apache 2.0) | **Not stated.** The README names no dataset and carries no acknowledgement; `data/` holds a class-ratio plot and nothing else. |
+| Input | one eye, 40×24, RGB, `x/255` | one head, 48×48, RGB, `x/255` |
+| Output | `prob_open`, a sigmoid | `prob_sunglasses`, a sigmoid |
+| sha256 as fetched | `9a346a08…c604ba9b64` | `9c13d937…a4063c73` |
+| sha256 as shipped | `c848d34c…4b29d04c7c` | `d49b6206…7e9538525c6` |
+
+Both shipped with a dynamic batch dimension that tract loads but the project pins anyway, with the
+same script as the pair:
+
+    ./tools/fix-face-model-shapes.sh ocec_s.onnx         models/face/ocec_s_b1.onnx   --dim batch=1
+    ./tools/fix-face-model-shapes.sh sgc_is_l_48x48.onnx models/face/sgc_l_48_b1.onnx --dim batch=1
+
+The one thing D13's discipline turns up here is SGC's undocumented training set. The weights'
+grant is MIT and that is what binds a redistributor; but "trained on what" is the question this
+project reads first, and for SGC it has no answer. Recorded so it is a known gap rather than an
+assumption, and so that whoever finds a sunglasses classifier with a stated dataset knows what to
+replace.
+
+Attribution, as ODC-By asks for the eye dataset:
+
+> Michał Młodawski, *Open and Closed Eyes Dataset*, July 2024,
+> https://huggingface.co/datasets/MichalMlodawski/closed-open-eyes
+
+The variant choice was measured, not taken from the F1 column: on family snapshots the S variant
+of OCEC read more open eyes as open than M or L did, which overfit their own domain
+(docs/faces.md §17).

packaging/PKGBUILD

@@ -61,7 +61,13 @@ package() {
     # These live in LFS; a checkout without `git lfs pull` has ~130-byte
     # pointers here. Installing one produces a package whose face indexing
     # fails inside the graph loader on the user's machine, so refuse instead.
-    for _m in scrfd_500m_640.onnx scrfd_2.5g_640.onnx scrfd_10g_640.onnx arcface_mbf_b1.onnx; do
+    #
+    # The last three are the eye-state models (docs/faces.md §17): landmarks,
+    # open/closed, sunglasses. Optional to the app, which indexes without
+    # them, but shipped beside the pair so the eyes-open filter works out of
+    # the box.
+    for _m in scrfd_500m_640.onnx scrfd_2.5g_640.onnx scrfd_10g_640.onnx arcface_mbf_b1.onnx \
+              2d106det_b1.onnx ocec_s_b1.onnx sgc_l_48_b1.onnx; do
         _src="models/face/${_m}"
         if [[ "$(stat -c%s "${_src}")" -lt 100000 ]]; then
             echo "error: ${_m} is an LFS pointer, not a model — run: git lfs pull" >&2

packaging/flatpak/paris.tourolle.darkroom.yml

@@ -157,7 +157,8 @@ modules:
       # inside the graph loader on the user's machine. Refuse instead, with the
       # command that fixes it.
       - |
-        for m in scrfd_500m_640.onnx scrfd_2.5g_640.onnx scrfd_10g_640.onnx arcface_mbf_b1.onnx; do
+        for m in scrfd_500m_640.onnx scrfd_2.5g_640.onnx scrfd_10g_640.onnx arcface_mbf_b1.onnx \
+                 2d106det_b1.onnx ocec_s_b1.onnx sgc_l_48_b1.onnx; do
           if [ "$(stat -c%s "models/face/$m")" -lt 100000 ]; then
             echo "error: $m is an LFS pointer, not a model — run: git lfs pull" >&2
             exit 1

packaging/windows/darkroom.nsi

@@ -70,7 +70,7 @@ Section "DarkRoom" SecMain
   File "${STAGE}\darkroom.exe"
   File "${STAGE}\LICENSE"
 
-  ; The seven model files (four face, three scene), beside the executable,
+  ; The ten model files (seven face, three scene), beside the executable,
   ; which is where `library::system_face_models_dirs` looks on Windows —
   ; last, after the user's own directories, exactly as /usr/share is on Linux.
   SetOutPath "$INSTDIR\models"

tools/fix-face-model-shapes.sh

@@ -11,6 +11,14 @@
 #   ... det_10g.onnx   scrfd_10g_640.onnx  --input input.1=1,3,640,640
 #   ... w600k_mbf.onnx arcface_mbf_b1.onnx --dim None=1
 #
+# And the three eye-state models, verified 2026-09-19 (docs/faces.md §17).
+# The landmark model's batch is the literal "None" like the embedder's; the
+# two classifiers' is a *named* dim_param "batch":
+#
+#   ... 2d106det.onnx         2d106det_b1.onnx  --dim None=1
+#   ... ocec_s.onnx           ocec_s_b1.onnx    --dim batch=1
+#   ... sgc_is_l_48x48.onnx   sgc_l_48_b1.onnx  --dim batch=1
+#
 # ## Why this exists
 #
 # The InsightFace exports declare dynamic input dimensions — SCRFD's H and W,