← Run Index
Run summary · docs/findings/p3-c2-bubble-attribution.toml §6 施策 #5 (refine.every 100->200)
re200-12500-on-r2
refine.every=200 (candidate) A/B arm B: Lego 12.5k full run repeat 2/3
partial pass
01
設定
- dataset
- /Users/otkrickey/dev/3dgs-workspace/datasets/nerf_synthetic/lego
- gt_convention
- Premultiplied
- iterations
- 12,500
- seed
- 42
- capacity
- 1,000,000 splats
- sh_degree
- 3
- loss
- L1Ssim
- lambda
- 0.200
- ssim
- window=7 sigma=1
- backend.backward
- Simd
- backend.loss_path
- Gpu
- backend.emit_pairs
- Original
- backend.forward
- F32
- backend.dispatch
- PerStage
- backend.radix_scan
- GpuHierarchical
- backend.radix_scatter
- SimdBallot
- backend.binning
- Resident
- backend.tile_sort
- GlobalRadix
- backend.submission
- Chained
- backend.target_cache
- true
- backend.shader_fast_math
- true
- os_build
- 26.6.2 (25G83)
- backend.ssim_clear
- HostMemset
- backend.raster_grad_clear
- HostMemset
- backend.backward_uniform
- HostBuffer
- backend.lazy_c3_scratch
- false
- backend.ssim_walk
- PerChannel
02
結果
Val PSNR
35.74dB
Wallclock
1m 33s
Final splats
267,959
Final loss
6.129592e-3
03
3D ビューワー
学習結果 PLY を .splat 形式に変換し、antimatter15/splat の WebGL ビューワーで対話表示。
04
推移
Loss curve — 反復ごとの training loss
Splat count — refine の発火と最終構成
05
中間ログ抜粋
| iter | 本 run loss |
|---|---|
| 1 | 5.9385e-1 |
| 1,000 | 6.9244e-2 |
| 2,000 | 3.4166e-2 |
| 3,000 | 1.5221e-2 |
| 4,000 | 9.8825e-3 |
| 5,000 | 8.0172e-3 |
| 6,000 | 7.2987e-3 |
| 7,000 | 6.9077e-3 |
| 8,000 | 6.6965e-3 |
| 9,000 | 6.5068e-3 |
| 10,000 | 6.3757e-3 |
| 11,000 | 6.2151e-3 |
| 12,000 | 6.1545e-3 |
| 12,500 | 6.1296e-3 |