Skip to content

T4A — Vision domain shift (single-layer / ResNet)

Paper evidence: main.pdf · Block findings

Lemma: D4 · Stack: pytorch Nuisance key: domain_shift

Production change: New camera, site, or geography; same classes.

Notebook (Run All, built-in demo): t04a-vision-domain.ipynb

pip install matching-pmh torch
# Open the notebook and Run All

What this task achieved (headline)

E1_multiscale Gram PMH on DomainNet real→sketch: 42.15% test acc (+3.31 pp vs B0 38.84%).

B0 E1 E1_multiscale
38.84% 39.34% 42.15%

Paper preset: t4_domain_d4 · from pmh.benchmark.presets import get_preset

Note: Notebook = single-hook D4 with class-aligned Gram when loaders are (x,y). Paper multiscale DomainNet: the paper (DomainNet multiscale).

Subtasks (paper)

DomainNet real→sketch

E1_multiscale +3.31 pp test acc.

Preset: t4_domain_d4

Per-layer TDI geometry

Domain Gram on hook features.

Preset: t4_domain_d4

B0 / E1 / E1_multiscale training

Preset: t4_domain_d4

Run with matching-pmh

from pmh import PMHTrainer, evaluate_robust_fit
# nuisance="domain_shift"

Do not use PMH when

New classes at deploy without relabeling.

Replace demo data with yours

Swap demo loaders for your train_loader, source_batches, target_batches, and deploy holdout. Hook the backbone before your task head.

← All 13 tasks · Quickstart