lars 4ee75d0042 Clamp n_sec classification label to K_MAX ...

Real data has steps with up to ~37 secondaries, but the n_sec head only
has K_MAX+1=16 classes. The unclamped label occasionally overflowed
cross_entropy's valid range and crashed CUDA training with
"unique_by_key: failed to synchronize: cudaErrorAssert". The
continuous secondary targets were already truncated to K_MAX slots;
only this label was missed.

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

2026-07-08 10:11:17 +02:00

..

test_analysis.py

WIP: energy-conservation PoC analysis/transforms updates

2026-07-01 13:58:44 +02:00

test_bump_dataset_version.py

Fix ruff, ty, and pytest failures; apply ruff format

2026-07-02 16:59:22 +02:00

test_cli_predict.py

Add --comment option to predict, recorded in YAML sidecar

2026-07-06 09:37:48 +02:00

test_config.py

Apply ruff format and document lint/type tooling in CLAUDE.md

2026-06-18 17:40:27 +02:00

test_create_root_files.py

Fix ruff, ty, and pytest failures; apply ruff format

2026-07-02 16:59:22 +02:00

test_dataset.py

Implement Phase 2: secondary particle prediction

2026-06-29 11:34:31 +02:00

test_dwarf.py

Unify dataset/tooling scripts into a single dwarf Typer CLI

2026-07-02 09:52:37 +02:00

test_flow.py

Implement Phase 2: secondary particle prediction

2026-06-29 11:34:31 +02:00

test_geometry.py

Add autoregressive shower rollout driver

2026-07-08 10:10:36 +02:00

test_loader.py

Add tooling for a versioned geant_steps dataset layout

2026-06-25 16:46:48 +02:00

test_network.py

Implement Phase 2: secondary particle prediction

2026-06-29 11:34:31 +02:00

test_phase2.py

Implement Phase 2: secondary particle prediction

2026-06-29 11:34:31 +02:00

test_rollout.py

Add autoregressive shower rollout driver

2026-07-08 10:10:36 +02:00

test_steps_to_parquet_parallel.py

Fix ruff, ty, and pytest failures; apply ruff format

2026-07-02 16:59:22 +02:00

test_steps_to_parquet.py

Merge energy-conservation-poc into phase2-secondary-prediction

2026-07-06 12:18:30 +02:00

test_transforms.py

Clamp n_sec classification label to K_MAX

2026-07-08 10:11:17 +02:00