lars 2358a75ee1 feat: add eval-cost benchmark — Geant4 reference vs surrogate rollout timing ...

Closes the roadmap's long-standing "no eval-latency number exists for any
configuration" gap. Instruments `giant rollout` to record per-physical-step
wall-clock cost in its YAML sidecar, adds a measured Geant4/miniCaloSim
per-step reference (giant/analysis/geant4_reference.py, from a 3-energy,
4-event-count-per-energy local benchmark), and wires both into a new
eval_cost_per_step PlotSpec in the giant analyze gallery.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

2026-08-31 11:52:20 +02:00

..

legacy

Bump ruff line-length to 120 and reformat

2026-08-12 13:33:09 +02:00

test_amp.py

Add bf16 autocast to the training loop (gitea #47)

2026-08-17 15:36:12 +02:00

test_analysis_reduce.py

feat(analysis): per-step secondary multiplicity plots

2026-08-26 14:13:18 +02:00

test_bump_dataset_version.py

Move scripts/ to giant/tools/ (issues.md Issue 9)

2026-08-13 10:31:47 +02:00

test_catalog.py

feat: add eval-cost benchmark — Geant4 reference vs surrogate rollout timing

2026-08-31 11:52:20 +02:00

test_checkpoint_io.py

Merge remote-tracking branch 'origin/master' into fix/issue-87

2026-08-28 11:31:05 +02:00

test_cli_new_run.py

Add per-stage init_from/freeze (gitea #42)

2026-08-17 14:23:20 +02:00

test_cli_predict.py

Merge remote-tracking branch 'origin/master' into fix/issue-87

2026-08-28 11:31:05 +02:00

test_cli_rollout.py

feat: add eval-cost benchmark — Geant4 reference vs surrogate rollout timing

2026-08-31 11:52:20 +02:00

test_cli_train_overrides.py

perf: defer heavy imports in giant/dwarf CLIs until commands run

2026-08-28 14:10:19 +02:00

test_cond_layout.py

Give the cond_cat/cond_cont column layout one owner (gitea #37)

2026-08-14 14:16:20 +02:00

test_condor.py

feat: add eval-cost benchmark — Geant4 reference vs surrogate rollout timing

2026-08-31 11:52:20 +02:00

test_config_consumed_keys.py

Implement stage2_model.stage1_context = "sampled" (gitea #41)

2026-08-17 12:27:13 +02:00

test_config.py

Add sampled n_sec under n_sec.mode = 'head' (gitea #86)

2026-08-28 11:04:45 +02:00

test_create_root_files.py

Move scripts/ to giant/tools/ (issues.md Issue 9)

2026-08-13 10:31:47 +02:00

test_dataset.py

Move scripts/ to giant/tools/ (issues.md Issue 9)

2026-08-13 10:31:47 +02:00

test_dwarf.py

Let dwarf warm-cache take --config so it can't under-warm a config's cache keys (gitea #59)

2026-08-17 08:54:14 +02:00

test_flow.py

Pass ConditioningAxisConfig/ParticleTypeConfig themselves instead of raw dicts (gitea #38)

2026-08-14 15:03:55 +02:00

test_geometry.py

Bump ruff line-length to 120 and reformat

2026-08-12 13:33:09 +02:00

test_layers.py

Deduplicate n_sec_head/type_head MLPs into build_mlp_head (gitea #36)

2026-08-14 10:57:47 +02:00

test_loader.py

Add class-balanced secondary particle-type loss (gitea #44)

2026-08-17 23:02:50 +02:00

test_materials.py

Bump ruff line-length to 120 and reformat

2026-08-12 13:33:09 +02:00

test_migration_v02_v03.py

Unify the two v0.2->v0.3 migration surfaces (issues.md Issue 6)

2026-08-13 09:56:22 +02:00

test_model_summary.py

Merge remote-tracking branch 'origin/master' into fix/issue-87

2026-08-28 11:31:05 +02:00

test_network.py

Give CriticModel a registry-built trunk and StageModel base (gitea #57)

2026-08-24 09:48:26 +02:00

test_objectives.py

Pass ConditioningAxisConfig/ParticleTypeConfig themselves instead of raw dicts (gitea #38)

2026-08-14 15:03:55 +02:00

test_particles.py

Bump ruff line-length to 120 and reformat

2026-08-12 13:33:09 +02:00

test_phase2.py

Clip raw predicted log_mass in decode_secondaries (gitea #54)

2026-08-17 09:31:37 +02:00

test_pipeline.py

Add class-balanced secondary particle-type loss (gitea #44)

2026-08-17 23:02:50 +02:00

test_release_tooling.py

Auto-bump patch version, tag, and update changelog on merge to master (gitea #50)

2026-08-18 10:40:19 +02:00

test_render.py

feat: add eval-cost benchmark — Geant4 reference vs surrogate rollout timing

2026-08-31 11:52:20 +02:00

test_rollout.py

Implement n_sec.mode = "stop_token" for the AR secondary decoder (gitea #40)

2026-08-17 10:44:14 +02:00

test_router_gating.py

Add multi-rollout support to giant analyze (gitea #77)

2026-08-24 13:23:50 +02:00

test_router.py

Add "none" variants for router, history, and trunk (gitea #45)

2026-08-24 09:22:36 +02:00

test_sample.py

Add sampled n_sec under n_sec.mode = 'head' (gitea #86)

2026-08-28 11:04:45 +02:00

test_setup_cache.py

Add class-balanced secondary particle-type loss (gitea #44)

2026-08-17 23:02:50 +02:00

test_steps_to_parquet_parallel.py

Move scripts/ to giant/tools/ (issues.md Issue 9)

2026-08-13 10:31:47 +02:00

test_steps_to_parquet.py

Move scripts/ to giant/tools/ (issues.md Issue 9)

2026-08-13 10:31:47 +02:00

test_train.py

Add class-balanced secondary particle-type loss (gitea #44)

2026-08-17 23:02:50 +02:00

test_training_plots.py

Fix LaTeX-unavailable skip check in analyze metrics smoke test

2026-08-24 11:15:51 +02:00

test_transforms.py

Give the cond_cat/cond_cont column layout one owner (gitea #37)

2026-08-14 14:16:20 +02:00

test_type_embedding_distance.py

Add multi-rollout support to giant analyze (gitea #77)

2026-08-24 13:23:50 +02:00

test_validate.py

Pass ConditioningAxisConfig/ParticleTypeConfig themselves instead of raw dicts (gitea #38)

2026-08-14 15:03:55 +02:00

test_wgan.py

Give CriticModel a registry-built trunk and StageModel base (gitea #57)

2026-08-24 09:48:26 +02:00