Demos

Canonical demo datasets — the same scenes used in the InverseNet paper: 10 KAIST-10 hyperspectral scenes for CASSI and the 6 standard SCI Video Benchmark videos (Kobe, Traffic, Runner, Drop, Crash, Aerial) for CACTI. Each sample is 256×256, deterministic, and committed to the repo so external users can verify reconstruction byte-for-byte. Download the input, run the reference solver locally, and see the PWM pipeline work end-to-end — or compete with a better solver.

T1_nominal

cacti

L3-004 (CACTI) — InverseNet SCI Video Benchmark

Full benchmark page → L3-004

reference PSNR

13.9 dB

your solver should beat this

Shape summary

snapshot: [256, 256]

ground truth: [8, 256, 256]

reference solution: [8, 256, 256]

sample_01 · kobe

ref PSNR 13.9 dB

Measurement

Ground truth

Reference

snapshot.npz352.8 KB solution.png42.6 KB meta.json1.2 KB ground_truth.npz1.72 MB ground_truth.png28.3 KB snapshot.png39.4 KB solution.npz1.14 MB

sample_02 · traffic

ref PSNR 8.42 dB

Measurement

Ground truth

Reference

snapshot.npz257.4 KB solution.png43.6 KB meta.json1.2 KB ground_truth.npz548.3 KB ground_truth.png38.0 KB snapshot.png47.5 KB solution.npz1.13 MB

sample_03 · runner

ref PSNR 11.86 dB

Measurement

Ground truth

Reference

snapshot.npz247.5 KB solution.png43.4 KB meta.json1.2 KB ground_truth.npz429.7 KB ground_truth.png30.6 KB snapshot.png38.0 KB solution.npz1.12 MB

sample_04 · drop

ref PSNR 4.69 dB

Measurement

Ground truth

Reference

snapshot.npz254.4 KB solution.png42.9 KB meta.json1.2 KB ground_truth.npz254.7 KB ground_truth.png16.4 KB snapshot.png36.9 KB solution.npz1.10 MB

sample_05 · crash

ref PSNR 7.34 dB

Measurement

Ground truth

Reference

snapshot.npz309.5 KB solution.png43.8 KB meta.json1.2 KB ground_truth.npz756.7 KB ground_truth.png28.8 KB snapshot.png41.6 KB solution.npz1.13 MB

sample_06 · aerial

ref PSNR 10.99 dB

Measurement

Ground truth

Reference

snapshot.npz331.1 KB solution.png44.6 KB meta.json1.2 KB ground_truth.npz835.0 KB ground_truth.png31.0 KB snapshot.png42.5 KB solution.npz1.12 MB

Run it

python3 pwm-team/pwm_product/reference_solvers/cacti/cacti_pnp_admm.py --input pwm-team/pwm_product/demos/cacti/sample_01 --output /tmp/out

T1_nominal

cassi

L3-003 (CASSI) — InverseNet KAIST-10 benchmark

Full benchmark page → L3-003

reference PSNR

19.9 dB

your solver should beat this

Shape summary

snapshot: [256, 256]

ground truth: [256, 256, 28]

reference solution: [256, 256, 28]

sample_01 · scene01

ref PSNR 19.86 dB

Measurement

Ground truth

Reference

snapshot.npz345.7 KB solution.png12.3 KB meta.json1.1 KB ground_truth.npz6.19 MB ground_truth.png23.4 KB snapshot.png34.9 KB solution.npz435.9 KB

sample_02 · scene02

ref PSNR 17.25 dB

Measurement

Ground truth

Reference

snapshot.npz345.3 KB solution.png15.6 KB meta.json1.1 KB ground_truth.npz6.30 MB ground_truth.png25.4 KB snapshot.png35.3 KB solution.npz440.1 KB

sample_03 · scene03

ref PSNR 12.34 dB

Measurement

Ground truth

Reference

snapshot.npz346.3 KB solution.png12.1 KB meta.json1.1 KB ground_truth.npz6.24 MB ground_truth.png17.5 KB snapshot.png37.8 KB solution.npz432.4 KB

sample_04 · scene04

ref PSNR 14.6 dB

Measurement

Ground truth

Reference

snapshot.npz344.0 KB solution.png9.1 KB meta.json1.1 KB ground_truth.npz6.22 MB ground_truth.png14.8 KB snapshot.png29.1 KB solution.npz432.1 KB

sample_05 · scene05

ref PSNR 15.47 dB

Measurement

Ground truth

Reference

snapshot.npz345.9 KB solution.png13.1 KB meta.json1.1 KB ground_truth.npz6.21 MB ground_truth.png19.3 KB snapshot.png37.0 KB solution.npz430.6 KB

sample_06 · scene06

ref PSNR 18.75 dB

Measurement

Ground truth

Reference

snapshot.npz346.4 KB solution.png11.5 KB meta.json1.1 KB ground_truth.npz6.22 MB ground_truth.png16.6 KB snapshot.png27.7 KB solution.npz441.3 KB

sample_07 · scene07

ref PSNR 13.46 dB

Measurement

Ground truth

Reference

snapshot.npz344.9 KB solution.png14.7 KB meta.json1.1 KB ground_truth.npz6.11 MB ground_truth.png28.3 KB snapshot.png41.0 KB solution.npz428.1 KB

sample_08 · scene08

ref PSNR 18.46 dB

Measurement

Ground truth

Reference

snapshot.npz346.3 KB solution.png11.5 KB meta.json1.1 KB ground_truth.npz6.28 MB ground_truth.png17.9 KB snapshot.png28.2 KB solution.npz443.5 KB

sample_09 · scene09

ref PSNR 11.66 dB

Measurement

Ground truth

Reference

snapshot.npz344.0 KB solution.png10.9 KB meta.json1.1 KB ground_truth.npz6.22 MB ground_truth.png17.7 KB snapshot.png37.3 KB solution.npz436.8 KB

sample_10 · scene10

ref PSNR 14.68 dB

Measurement

Ground truth

Reference

snapshot.npz346.9 KB solution.png12.1 KB meta.json1.1 KB ground_truth.npz6.17 MB ground_truth.png22.8 KB snapshot.png32.6 KB solution.npz439.7 KB

Run it

python3 pwm-team/pwm_product/reference_solvers/cassi/cassi_gap_tv.py --input pwm-team/pwm_product/demos/cassi/sample_01 --output /tmp/out

Demo files are regenerated from scripts/generate_demos.py at fixed RNG seeds — byte-stable across runs so any SHA-256 in a sample's meta.json can be cross-checked against the repo.