Photon Memory checkpoints (Rough Flower / PureLight comparison)

Companion weights for the Photon Memory reader that matches / beats PureLight on directional NLL for the released Rough Flower luminaire. Code: ANABUR920/renderTTT (lightmem/).

Files

Path Role Size
checkpoints/v3_ksched_model.pt Meta-trained PhotonKernel reader (k_feat=True) ~0.7 MB
checkpoints/big_1e8_k512_ttt.pt Reader-only TTT on 1e8 memory (k=512) ~0.7 MB
checkpoints/big_match_res2b_best.pt Intermediate matched TTT + residual (~NLL βˆ’4.54) ~42 MB
checkpoints/big_match_res2c_best.pt Final matched TTT + residual (NLL βˆ’6.00, render default) ~42 MB
logs/*.jsonl Training / eval logs for the above TTT runs small

Download

hf download ANABUR/photon-memory --local-dir ./photon-memory-ckpts
# Photons (separate dataset):
hf download ANABUR/photon-memory-photons --repo-type dataset --local-dir ./photon-memory-photons

Reproduce the βˆ’6.00 NLL claim (same protocol as paper)

Needs: CUDA GPU (~40 GB for 1e8 + residual), traced Rough Flower photons (flower_surface_* 11e6 held-out + flowerbig_surface_* 1e8 memory), and this repo's scripts/ttt_big.py.

Winning warm-start chain (do not claim a cold start from the meta reader equals βˆ’6.00 in one shot):

  1. Start from v3_ksched_model.pt β†’ reader-only TTT β†’ big_1e8_k512_ttt.pt (~βˆ’0.53).
  2. Joint reader + residual matched TTT β†’ big_match_res2b_best.pt (~βˆ’4.54).
  3. Continue:
python scripts/ttt_big.py \
  --data flowerbig_surface --N 100000000 --k 512 --wmult 2 \
  --ttt_matched --n_ttt_q 1000000 --steps 8000 --residual \
  --res_K 32 --res_log2 19 --res_finest 2048 \
  --init_ttt checkpoints/big_match_res2b_best.pt \
  --lr 5e-6 --res_lr_grid 5e-3 --res_lr_head 5e-4 \
  --name big_match_res2c

Zero-shot (no TTT) with v3_ksched_model.pt at N=1e8, k=512 reaches RGB-mean NLL β‰ˆ βˆ’0.460, matching the PureLight teacher plateau (~βˆ’0.46). The βˆ’6.00 number is matched test-time adaptation on the non-eval remainder of the same 11e6 photon file as the held-out set (memory may be a separately traced 1e8 bank). Leave-one-out TTT on the 1e8 memory alone stalls near βˆ’0.53. It is not zero-shot and not a fresh-retrace claim.

Honest notes for reviewers / friends

  • Held-out queries are always the last 20k of default_rng(0).permutation(11_000_000) on flower_surface_*.
  • --ttt_matched trains on the non-eval remainder of that same 11e6 file while retrieving from the separately traced 1e8 memory.
  • A frozen-reader hash residual alone peaks ~βˆ’0.38 and does not beat PureLight.
  • Photons for Rough Flower are published separately: ANABUR/photon-memory-photons (flower_surface_* 11e6 + flowerbig_surface_* 1e8, channels mean/r/g/b). PureLight teacher weights and photon maps remain third-party; otherwise re-trace with scripts/gen_flower.py.

Photons (required for reproduction)

Rough Flower OptiX traces live in ANABUR/photon-memory-photons:

hf download ANABUR/photon-memory-photons --repo-type dataset --local-dir ./photon-memory-photons

Layout: flower_surface_{mean,r,g,b}/val/rough_flower.npz (11e6) and flowerbig_surface_{mean,r,g,b}/val/rough_flower.npz (1e8).

Load in code

import torch
from lightmem.models import PhotonKernel, ResidualLobes

sd = torch.load("checkpoints/big_match_res2c_best.pt", map_location="cpu")
reader = PhotonKernel(k=512, k_feat=True)
reader.load_state_dict(sd["reader"])
residual = ResidualLobes(K=32, finest=2048, log2_table=19)
residual.load_state_dict(sd["residual"])
# sd["nll"] β‰ˆ -6.00, sd["step"] == 8000
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