Adinfinitium

Adinfinitium Terrain Diffusion (experimental)

Status: Experimental Β· License: Apache-2.0 Β· Domain: lunar south pole (NASA LRO LOLA 5 m)

Conditional diffusion U-Net that fills terrain height between laser points on 64Γ—64-pixel patches. Experimental: classical kriging scores better, so kriging is the estimator we ship. Outputs are always labelled GENERATED – NOT OBSERVED.

Model details

Task Terrain-height gap filling (hide-and-check)
Parameters 49,943,681
Architecture DenoiseAttnUNet: conditional denoiser, residual blocks, time embedding, self-attention at coarse scales. base 128, depth 3, T 100, blocks 2, heads 8, attn_levels 2
Training 7,671 of 8,000 steps; best = step 7,500; trained only on laser-measured pixels by hide-and-check
Checkpoint terrain_diffusion_mini.pt Β· 199,839,435 bytes Β· id 1f647a8660 Β· sha256 1a8bfbad…0e643
Wiring Read by nothing downstream

Benchmarks

Kriging scores 0.97 m RMSE and the network 2.01 m on the held-out sites, so kriging is the shipped estimator and this model is an experiment.

Terrain: kriging vs diffusion network

Hide-and-check RMSE on hidden laser heights (m, lower is better) Kriging Diffusion network Old gap-fill backup
Site23 0.94 1.83 17.92
Site42 1.11 2.80 24.82
Six fresh sites 0.97 1.62 n/a

Uncertainty band: the raw 5–95% band fails (21.9% / 37.3% coverage on Site23 / Site42). One split-conformal scale (8.232, fit on Site11 only) gives 84.9% / 96.9% coverage, but the band is wide (7.06 m / 15.51 m).

Limitations

  • Seven of nine classical methods, including plain linear interpolation, beat the network on both held-out sites; a fitted combination puts weight 0.0064 on it.
  • The test favours interpolation because hidden pixels sit next to shown ones; wide gaps were never tested for either method.
  • The calibrated band is wide and uneven (6 of 19 windows outside 80–97% coverage), with no formal guarantee.

Usage

# project code: https://github.com/oasisk/adinfinitium-handoff
from worldmodel.terrain.model import load_checkpoint
from worldmodel.terrain.generate import generate_versions

Dream-RSI: gated offline self-improvement

Dream-RSI lets a model keep improving after the main training ends, offline and safely:

  1. Find failures: fly or replay many missions and record where the current model fails.
  2. Dream harder: generate and replay those hard cases (failure-boosted scenarios) and retrain a candidate, warm-started from the current model.
  3. Gate: the candidate is scored on a frozen validation exam (fixed seeds, paired with the current model and the physics backup).
    • Lander: accept only if the candidate loses no mission that the current model or the physics backup lands (zero regressions) AND improves total success.
    • Planner and rover: accept only if the frozen validation metric improves by a margin.
  4. Repeat from whichever model won. Rejected candidates are discarded, and the shipped model never gets worse.

Test maps and the final-exam site are never used in Dream-RSI: training uses training maps, and gating uses a separate validation map.

Measured results (all logged rounds)

Dream-RSI rounds accepted per loop

Loop Rounds accepted Effect on its gate metric
Planner (PlannerSetNet) 1 of 5 Validation Brier 0.01543 β†’ 0.01353 (βˆ’12.3%)
Lander pilot (Site11 gate, 1,024 paired missions) 0 of 7 Pilot unchanged; every candidate lost 12–25 missions, so the gate held
Rover brain v3.1 (self-learning on stressed worlds) 9 of 36 Blocked route improved at 4/4 sites, late light 3/4, dusty panels 1/4, early light 0/4
Rover brain v3 9 of 72 Blocked route improved at 6/6 sites, dusty panels 2/6
Rover learned driver v2 5 of 38 Returned rate stayed 0.0%

Across all loops, 18 of 42 first rounds on a new problem were accepted, against 6 of 120 later rounds. What this shows: Dream-RSI reliably repairs specific failures (blocked routes, late light, model faults) without ever shipping a regression. In unchanged conditions it rarely finds a safe gain (0 of 32 rover rounds accepted).

Projections after more Dream-RSI (Projected, not measured)

Median after 20 more generations, from the planner's own Dream-RSI log:

  • Planner Brier: 0.0145 β†’ 0.0094 [0.0043, 0.0145]

Scenario only, for models where Dream-RSI has not been run yet (it borrows the improvement rate seen in the planner and rover loops):

  • Hazard fresh F1: 0.757 β†’ 0.779
  • Light error: 0.450Β° β†’ 0.410Β°
  • Terrain error: 2.01 β†’ 1.84 m

The lander and rover v3.1 are projected to stay unchanged, because their gates have not accepted anything in unchanged conditions.

Limitations and next steps

  • The zero-regression gate is strict: it also rejected lander candidates that were slightly better overall (best: 991 vs 990 landings).
  • Training rounds were short (4–5 updates) because of the compute budget.
  • Next: harder self-generated scenarios, longer rounds, and a fresh unseen-map exam after each accepted generation.

Intended use and scope

Research, education and benchmarking of lunar landing-site navigation methods. Not for flight, safety-critical or operational landing decisions. All results are simulation or held-out-lunar-site results, not real-mission performance. Models were frozen (SHA-256) before held-out evaluation; each test site was evaluated once.

About Adinfinitium

Adinfinitium is an open-source effort to make lunar landing and surface navigation easier to study. It began as our NASA Space Apps 2026 project, Lunar Landing Site Navigator, and trains on NASA LRO LOLA 5 m terrain of the lunar south pole. The model chain is: hazard network β†’ light network β†’ terrain model β†’ rover brain β†’ lander pilot. See the whole family at https://huggingface.co/Imok282.

Source code, full records and the hand-off package: https://github.com/oasisk/adinfinitium-handoff (checkpoint SHA-256 values: CHECKPOINTS.sha256).

Citation

@misc{adinfinitium2026lunar,
  title  = {Adinfinitium open models for lunar landing-site navigation},
  author = {Adinfinitium},
  year   = {2026},
  howpublished = {Hugging Face, https://huggingface.co/Imok282},
  note   = {NASA Space Apps 2026, Lunar Landing Site Navigator}
}
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support