{ "bench_version": 1, "track": "euler_morphology_denoising", "model": "small_denoiser", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.006 }, "sweep": [ { "cfg": { "lr": 0.001 }, "mean": 0.2358667068183422 }, { "cfg": { "lr": 0.003 }, "mean": 0.19127917662262917 }, { "cfg": { "lr": 0.006 }, "mean": 0.11789883300662041 } ], "full": { "mean": 0.11414278019219637, "std": 0.019868887010459576, "per_seed": [ 0.11202351003885269, 0.09799353033304214, 0.13151243329048157, 0.13006585836410522, 0.1398763358592987, 0.08686374127864838, 0.12790271639823914, 0.0869041159749031 ], "n": 8 } }, "idea": { "mean": 0.1143862335011363, "std": 0.01959046199703673, "per_seed": [ 0.11207777261734009, 0.09821636974811554, 0.1314822882413864, 0.1300397664308548, 0.13983863592147827, 0.08825309574604034, 0.12812229990959167, 0.0870596393942833 ], "n": 8 }, "comparison": { "delta_mean": 0.00024345330893993378, "idea_wins": 3, "n_pairs": 8, "per_seed_diffs": [ 5.426257848739624e-05, 0.00022283941507339478, -3.014504909515381e-05, -2.6091933250427246e-05, -3.769993782043457e-05, 0.0013893544673919678, 0.00021958351135253906, 0.000155523419380188 ], "p_value": 0.0917, "mde": 0.0003977750936249918, "mde_rel_pct": 0.3484890528820207, "verdict": "no measurable effect", "system_worked": false }, "mechanism_signature": { "mechanism_signature": { "temperature": 0.1, "lambda": 0.2, "test_soft_mae": 0.04664662852883339, "test_hard_mae": 0.00214973883703351, "soft_pred_curve": [ 0.0162891186773777, 0.1149718388915062, 0.028045227751135826, 0.0041561247780919075 ], "soft_real_curve": [ 0.07483800500631332, 0.015983328223228455, 0.0032997075468301773, -0.00014746440865565091 ], "confirmed": true }, "custom_track": { "name": "euler_morphology_denoising", "file": "euler_custom_track.py", "domain": "vision" } }, "idea_sweep": [ { "cfg": { "lr": 0.001 }, "result": { "mean": 0.2288948018103838, "std": 0.011050052990791315, "per_seed": [ 0.23519256711006165, 0.2334323674440384, 0.24315214157104492, 0.23169168829917908, 0.2391384392976761, 0.2086285650730133, 0.22384829819202423, 0.21607434749603271 ], "n": 8 } }, { "cfg": { "lr": 0.003 }, "result": { "mean": 0.18089030869305134, "std": 0.021725469329973896, "per_seed": [ 0.19138242304325104, 0.17021414637565613, 0.20749393105506897, 0.1961212158203125, 0.20285634696483612, 0.13692161440849304, 0.17236995697021484, 0.16976283490657806 ], "n": 8 } }, { "cfg": { "lr": 0.006 }, "result": { "mean": 0.1143862335011363, "std": 0.01959046199703673, "per_seed": [ 0.11207777261734009, 0.09821636974811554, 0.1314822882413864, 0.1300397664308548, 0.13983863592147827, 0.08825309574604034, 0.12812229990959167, 0.0870596393942833 ], "n": 8 } } ], "protocol_notes": { "structural_match": "image-field denoising with spatial excursion topology", "epochs": 5, "batch": 128, "baseline_and_idea_lr_union": [ 0.001, 0.003, 0.006 ], "baseline_selection_seeds": [ 0, 1, 2, 3 ], "paired_seeds": [ 0, 1, 2, 3, 4, 5, 6, 7 ] } }