Re-export SpatialLM glb files with lehoangan02/SpatialLM 7021237 (close wall gaps; fixes P63 main room)

Bedroom1/Spatiallm1_MasterSlam1/room_with_scan_preview.glb

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2c06a4998f996f0a0dd64130315aafc50e7097fca3614b477442065847e6a2b6
+oid sha256:1b19b754b1c7ececeae534d42b07d673de97c8abdb9f7468f08a713829c33189
 size 33622412

I87X/Spatiallm1_MasterSlam1/room_with_scan_preview.glb

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1f47132d9c889bf7bf439dfcdee10f89ab6957896b7517ad4d2f7567e96b2a34
+oid sha256:473d650053186f572c9900296e9c8f20bfcaa79fb14cc7978c2c540841f12b2e
 size 33647392

I87X/Spatiallm1_MasterSlam2/room_with_scan_preview.glb

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1cb8d93f570e9e4ea1aaeb4236c1482d84821c85038346ba97c9ddcdb473490d
+oid sha256:c0d454952e941deebd5909499616917f170ecab5ab9affd4a14ad7b6cb41212f
 size 33617864

P63/Spatiallm1_MasterSlam1/room_with_scan.glb

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:115514bc313ba71bf527102252eddeeb4c3ec54470f83b82f64f4b378674f77a
-size 4819720
+oid sha256:3c144fa86f2a1507429797258cdd7a32549279995b856ca66f024db70f018bc5
+size 4821880

P63/Spatiallm1_MasterSlam1/room_with_scan_preview.glb

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:329b6ea3d7c25fa0fd73d33edf6ca61a2cdefe5702bf1d7fe9ea05ae8f6bcbdd
-size 33620024
+oid sha256:a1a93c07eb9a71dcc313365b2c87c8d249a17a8357a56e85e4c11c7a1165b6f9
+size 33622184

P64/Spatiallm1_MasterSlam1/room_with_scan_preview.glb

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:63fb3283a8948fc5743b02f63745d43da10e18c4ad2db20c3fcce85f2adeef9a
+oid sha256:ab732fe48738e924cceef58a9d74e20c6bfdc17d22f996e113cd2e62c1d1a9ad
 size 33618668

README.md

@@ -81,7 +81,7 @@ A trailing number is a run/capture index within the room.
 | [MASt3R-SLAM](https://github.com/rmurai0610/MASt3R-SLAM) | Dense SLAM point cloud and camera trajectory from video (Colab or a GPU server; uncalibrated unless noted in Run notes) | `MasterSlam*_*` |
 | `room_plane_fit.py` (own script, in the folder) | Cleans the point cloud, aligns to gravity, fits wall planes, makes mesh, colliders and walls (Unity-ready) | `MasterSlam1_ScanVideo1/room_plane_fit_out/` |
 | `ply2obj.py` (own script) | Converts `scan.ply` to a textured mesh (`--mesh`: Poisson + baked texture) | `scan_mesh.obj/.mtl/.png` |
-| [SpatialLM](https://github.com/manycore-research/SpatialLM), run with the fork [lehoangan02/SpatialLM](https://github.com/lehoangan02/SpatialLM) (`stage3/pipeline.py`, commit `a10e67a`) | Room layout (walls, objects) from a point cloud | `Spatiallm1_*/room.glb`, `room_with_scan.glb` |
+| [SpatialLM](https://github.com/manycore-research/SpatialLM), run with the fork [lehoangan02/SpatialLM](https://github.com/lehoangan02/SpatialLM) (`stage3/pipeline.py`, commit `a10e67a`; `.glb` files re-exported with `7021237`) | Room layout (walls, objects) from a point cloud | `Spatiallm1_*/room.glb`, `room_with_scan.glb` |
 | NeuCon | Mesh reconstruction (trimesh export) | `Bedroom1/NeuCon1/neucon.ply` |
 | RTAB-Map | RGB-D mapping, textured mesh | `P64/RTABMAP1/*.obj/.mtl/.jpg` |
 
@@ -142,7 +142,7 @@ scene for Preview / Quick Look: each scan point is a 1.5 cm coloured tetrahedron
   with the ARCore intrinsics from `Video2_T3/intrinsics.json` (`--calib`). Input: the 1870 NV21
   frames converted to PNG (1920x1080). Against the ARCore poses (Sim3-aligned), both have ~15 cm
   ATE; the calibrated run has about half the rotation error between keyframes (0.60 vs 1.16 deg).
-- All `Spatiallm1_*` folders: fork `stage3/pipeline.py` (commit `a10e67a`), Llama-1B, seeds 0 1 2, default
+- All `Spatiallm1_*` folders: fork `stage3/pipeline.py` (commit `a10e67a`; `.glb` re-exported with `7021237`, which closes wall gaps under 20 cm), Llama-1B, seeds 0 1 2, default
   2.6 m ceiling height (sets the scale). Non-default options, recorded in `summary.json`:
   - `P64/Spatiallm1_MasterSlam1`: `--up_hint=-y`. With `--up_hint auto` the wrong axis was taken
     as up (0.085 raw height, scale x30). Only seed 0 found walls.