# tag=m1-candle start=2026-11-03T17:53:11+01:00 cmd=<home>/supratex-0311/venv/bin/python scripts/run_example.py --config <home>/supratex-0311/ready/models.json --input <home>/supratex-0311/SUPRA-Tex/assets/metalwork_candle_jar.png --output <home>/supratex-0311/out/m1-candle.glb
[    0.62] [Supra-Tex inference] attention=sdpa; sparse_attention=sdpa; low_vram=False
[    0.62] [1/2] Foreground, MoGe-2, geometry, material maps -> <home>/supratex-0311/out/m1-candle_work/input
[    0.62] [Supra-Tex inference] attention=sdpa; sparse_attention=sdpa; low_vram=False
[   97.11] [2/2] Surface observations, PBR completion, GLB export -> <home>/supratex-0311/out/m1-candle.glb
[   97.11] [Supra-Tex inference] attention=sdpa; sparse_attention=sdpa; low_vram=False
[   97.41] [SPARSE] Conv backend: flex_gemm; Attention backend: sdpa; Batch broadcast: direct
[   97.41] Floaters: removed 0 of 2 components (0 of 21120746 faces)
[  127.72] Observation: silhouette IoU 0.984 (geometry inside image 1.000, image covered 0.984); admitted 641095 of 641608 trusted pixels (missed geometry 46, depth edge 87, image-depth mismatch 380)
[  127.72] Image depth gate: applied; 647958 paired depth pixels, 855 mismatch pixels
[  127.74] Observation image depth: excluded 171 additional voxel samples crossing the mismatch mask; 545749 observed voxels
[  131.79] [ATTENTION] Using backend: sdpa
[  131.79] Observation cache: enabled
[  167.51] 
Extracting GLB:   0%|          | 0/6 [00:00<?, ?it/s]
Building BVH:  17%|█▋        | 1/6 [00:00<00:00, 48.89it/s]
Building BVH:  33%|███▎      | 2/6 [00:05<00:11,  2.84s/it]
Cleaning mesh:  33%|███▎      | 2/6 [00:05<00:11,  2.84s/it]<home>/supratex-0311/venv/lib/python3.12/site-packages/cumesh/remeshing.py:220: UserWarning: Using torch.cross without specifying the dim arg is deprecated.
[  167.51] Please either pass the dim explicitly or simply use torch.linalg.cross.
[  167.51] The default value of dim will change to agree with that of linalg.cross in a future release. (Triggered internally at /__w/pytorch/pytorch/aten/src/ATen/native/Cross.cpp:63.)
[  167.51]   normals0 = torch.cross(mesh_vertices[atempt_triangles_0[:, 1]] - mesh_vertices[atempt_triangles_0[:, 0]], mesh_vertices[atempt_triangles_0[:, 2]] - mesh_vertices[atempt_triangles_0[:, 0]])
[  205.07] 
Cleaning mesh:  50%|█████     | 3/6 [00:06<00:05,  1.99s/it]
Parameterizing new mesh:  50%|█████     | 3/6 [00:06<00:05,  1.99s/it]
Parameterizing new mesh:  67%|██████▋   | 4/6 [00:36<00:24, 12.30s/it]
Sampling attributes:  67%|██████▋   | 4/6 [00:36<00:24, 12.30s/it]    
Sampling attributes:  83%|████████▎ | 5/6 [00:37<00:08,  8.60s/it]
Finalizing mesh:  83%|████████▎ | 5/6 [00:37<00:08,  8.60s/it]    
Finalizing mesh: 100%|██████████| 6/6 [00:43<00:00,  7.62s/it]
Finalizing mesh: 100%|██████████| 6/6 [00:43<00:00,  7.23s/it]
[  209.53] Textured mesh: <home>/supratex-0311/out/m1-candle.glb
[  209.53] Intermediate files and run receipt: <home>/supratex-0311/out/m1-candle_work