diff --git a/artifacts/s2-model-synthetic-2026-11-14/receipt.json b/artifacts/s2-model-synthetic-2026-11-14/receipt.json new file mode 100644 index 000000000..0b02e6260 --- /dev/null +++ b/artifacts/s2-model-synthetic-2026-11-14/receipt.json @@ -0,0 +1,157 @@ +{ + "timestamp": "2026-11-14T14:44:43.272858+00:00", + "status": "PASS", + "returncode": 0, + "test_count": 20, + "test_child_cpu_seconds": 5.0396279999999996, + "test_child_peak_rss_kib": 567628, + "versions": { + "numpy": "1.26.4", + "scipy": "1.15.3", + "sklearn": "1.7.2", + "torch": "2.5.1+cu124" + }, + "cuda_visible_devices": "", + "torch_threads": 1, + "real_dataset_reads": false, + "fits": 0, + "backward_passes": 0, + "parameter_updates": 0, + "parameter_counts": { + "N-128": { + "encoder": 373408, + "decoder": 33409, + "total": 406817 + }, + "E-128": { + "encoder": 373408, + "decoder": 33921, + "total": 407329 + }, + "N-256": { + "encoder": 1473696, + "decoder": 132353, + "total": 1606049 + }, + "E-256": { + "encoder": 1473696, + "decoder": 133377, + "total": 1607073 + } + }, + "grid": [ + { + "learning_rate": 0.05, + "max_leaf_nodes": 31, + "max_iter": 200, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.05, + "max_leaf_nodes": 31, + "max_iter": 500, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.05, + "max_leaf_nodes": 31, + "max_iter": 1000, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.05, + "max_leaf_nodes": 63, + "max_iter": 200, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.05, + "max_leaf_nodes": 63, + "max_iter": 500, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.05, + "max_leaf_nodes": 63, + "max_iter": 1000, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.1, + "max_leaf_nodes": 31, + "max_iter": 200, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.1, + "max_leaf_nodes": 31, + "max_iter": 500, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.1, + "max_leaf_nodes": 31, + "max_iter": 1000, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.1, + "max_leaf_nodes": 63, + "max_iter": 200, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.1, + "max_leaf_nodes": 63, + "max_iter": 500, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + }, + { + "learning_rate": 0.1, + "max_leaf_nodes": 63, + "max_iter": 1000, + "min_samples_leaf": 20, + "random_state": 0, + "early_stopping": false + } + ], + "sha256": { + "tools/s2_models/__init__.py": "a0fa1323d3553172da66c5d442c5fc4fcb0d5055c886c776928604550ea97609", + "tools/s2_models/attention.py": "7714adfccf6a0fc6b52a44aee3a50321a88898bf3d743680a1108c01e1ff8a61", + "tools/s2_models/blocks.py": "b13e986cea5f1d7f325a1638c1505197b4a105baacff5a9a515f23fcaf28ea75", + "tools/s2_models/features.py": "760e580c811284876e592138931563ddddb7e0ca7c4a2fe8151705510d35e035", + "tools/s2_models/network.py": "cfff6c519b6d71f6c71b3bdf330e0fe63691f8d1a308c0e5e46ce7307def2697", + "tools/s2_models/objectives.py": "5b6527a45f08690055d572335470ce5979e8bb86e1cab00de594f79ae21316b1", + "tools/s2_models/requirements-tested.txt": "d0397e13dc138c303e4760eb93c52a28d2d260fc2eb91b12f119bf1da1f0463c", + "tools/s2_models/upstream/LICENSE": "2c919cd03fa823bf7eefc00a957ff8324c865cd22aa5285e563dc4b558084f25", + "tools/s2_models/upstream/Physics_Attention.py": "4fc710f6f96788fb59cb19fab8d64873d3d16e9e182f12a544d451c74c548cc7", + "tools/s2_models/upstream/Transolver_Irregular_Mesh.py": "5259759f7ecc6a954c8442ac8359990b9f242f1437c452c69e5e0f03549b5311", + "tools/s2_models/upstream/manifest.json": "fbcaa6673dbc82031e110ea18680574a9d9edcbf33ca6abd537129be4d860eee", + "tools/s2_models/upstream-adaptation.diff": "e62a4a7c24bfbc318365a84f4638f875bce587abb082738360bc7f12187c46c3", + "tools/test_s2_models.py": "f1c40626ee5e8f1b210ca7f7b665237f5809da920c1c0148e35b499ecf21ca6a" + }, + "acceptance_and_count_wall_seconds": 8.752674713963643, + "measurement_note": "Wall includes the test subprocess, parent library import and four model parameter-count constructions. Child CPU/RSS cover the unittest subprocess including imports; unittest internal duration is in unittest.txt. Data/training flags describe audited test/code scope, not an OS file-access trace." +} diff --git a/artifacts/s2-model-synthetic-2026-11-14/registration-monitor.json b/artifacts/s2-model-synthetic-2026-11-14/registration-monitor.json new file mode 100644 index 000000000..60404c51b --- /dev/null +++ b/artifacts/s2-model-synthetic-2026-11-14/registration-monitor.json @@ -0,0 +1,33 @@ +{ + "timestamp": "2026-11-14T14:44:43.272858+00:00", + "progress": { + "completed_jobs": 425, + "active_jobs": [ + 417, + 418, + 422, + 423, + 426, + 430, + 431, + 432, + 433, + 434, + 435, + 436, + 437, + 438, + 439, + 440 + ], + "next_job": 441, + "completed_worker_cpu_seconds": 35366.68382500002, + "live_tree_cpu_seconds": 801.2499999999997, + "active_tree_rss_bytes": 2344103936, + "stop_reason": null, + "elapsed_seconds": 2384.1496831209515 + }, + "terminal_receipts": 425, + "failed_receipts": 0, + "completion_flags": [] +} diff --git a/artifacts/s2-model-synthetic-2026-11-14/unittest.txt b/artifacts/s2-model-synthetic-2026-11-14/unittest.txt new file mode 100644 index 000000000..13f710012 --- /dev/null +++ b/artifacts/s2-model-synthetic-2026-11-14/unittest.txt @@ -0,0 +1,25 @@ +test_affine_p1_and_zero_gradient (__main__.GeometryTests) ... ok +test_base_slots_and_nonfinite_fail (__main__.GeometryTests) ... ok +test_degenerate_and_unused_fail (__main__.GeometryTests) ... ok +test_features_and_physical_scaling (__main__.GeometryTests) ... ok +test_moment_unequal_volumes_and_coincident_center (__main__.GeometryTests) ... ok +test_moment_volume_weight_rotation_and_filter (__main__.GeometryTests) ... ok +test_strict_volume_front_ties (__main__.GeometryTests) ... ok +test_unreachable_fail_closed (__main__.GeometryTests) ... ok +test_volume_lumping (__main__.GeometryTests) ... ok +test_wall_edges_removed_both_distances_and_gate_tie (__main__.GeometryTests) ... ok +test_four_models_forward_only_and_parameter_counts (__main__.NetworkTests) ... ok +test_original_attention_numerical_parity (__main__.NetworkTests) ... ok +test_query_order_global_context_and_input_rejection (__main__.NetworkTests) ... ok +test_upstream_snapshot_hashes (__main__.NetworkTests) ... ok +test_hierarchy_and_draw_compression (__main__.ObjectiveTests) ... ok +test_histgbrt_grid_unfitted (__main__.ObjectiveTests) ... ok +test_invalid_hierarchy (__main__.ObjectiveTests) ... ok +test_process_target_roundtrip_and_scale_invariance (__main__.ObjectiveTests) ... ok +test_score_masks_before_arithmetic_and_zero_truth (__main__.ObjectiveTests) ... ok +test_zero_missing_scales_no_fallback (__main__.ObjectiveTests) ... ok + +---------------------------------------------------------------------- +Ran 20 tests in 0.214s + +OK diff --git a/docs/s2-model-synthetic-contract-2026-11-14.md b/docs/s2-model-synthetic-contract-2026-11-14.md new file mode 100644 index 000000000..5c7371009 --- /dev/null +++ b/docs/s2-model-synthetic-contract-2026-11-14.md @@ -0,0 +1,58 @@ +# S2 R2-G24 / N / E synthetic code acceptance + +## Scope and decisions + +This implements Claude's [22:18:01 ruling](https://arcticnimbus.slack.com/archives/C0C78UFKGRM/p1794665881295239), the v0.5 Gre24 mapping and v0.6 process-conditioned target, with the [22:37:54 correction](https://arcticnimbus.slack.com/archives/C0C78UFKGRM/p1794667074550589). It is a synthetic-only code delivery for Claude's primary review, not a fitted model, real-corpus preprocessing receipt, or approval of predictive accuracy. + +No real geometry, LOG or simulation field is loaded by these modules/tests. The existing registration queue and its pinned sources are untouched. Test execution has empty CUDA visibility, one numerical thread and CPU tensors. No estimator `.fit`, backward pass, optimiser or parameter update is executed. Model initialisation is random, fixed by test seeds. + +**The latest target ruling replaces v0.3's transform entirely:** predict `t* = T/t_set` directly, loss on `t*`, seconds = prediction × `t_set`. There is no log transform, tau, fitted scale, oracle end-of-fill time or fallback. The initial pre-ruling tests were superseded; the receipt in this delivery tests the corrected implementation. `t_set` must be a finite positive process setting supplied equally to all arms. `B0 = V*_f` in normalised units; multiply by `t_set` for seconds. Predictions are not silently clipped. + +## R2-G24 computational contract + +`tools/s2_models/features.py` takes in-memory arrays; it has no data-loading entrypoint. + +- Coordinates are physical 3D positions; tetrahedra index all tet nodes in ascending original-ID order. Invalid indices, unused nodes, degenerate cells and nonfinite inputs fail closed. Exact `D_w=d(x)` is supplied by the existing distance preparation, not recomputed or inferred here. +- Tet edges are unique unordered pairs. Gate source is the nearest node to the supplied gate, with exact-distance ties going to the first/smaller original ID. +- Both R2 `D_i` and `D_f` exclude edges whose endpoints both have `D_w=0`. Remaining costs are length and length/mean(endpoint `D_w`), respectively. Unreachable nodes fail the part; no epsilon speed or imputation. This is a graph approximation of the eikonal, not tetFIM. +- Lumped volume is `Σ incident tet volume/4`. `V*_f(i)` is the strict weighted CDF `Σ_{D_f(j)