Integrated-Growth Hopf Delay Scheduler / bench_report.json

Failed on benchmark

Raw ⬇ ZIP
  1{
  2  "bench_version": 1,
  3  "track": "dynamics",
  4  "model": "rnn_small",
  5  "metric_direction": "lower is better",
  6  "n_seeds": 8,
  7  "baseline": {
  8    "best_cfg": {
  9      "lr": 0.006
 10    },
 11    "sweep": [
 12      {
 13        "cfg": {
 14          "lr": 0.001
 15        },
 16        "mean": 0.47112278640270233
 17      },
 18      {
 19        "cfg": {
 20          "lr": 0.003
 21        },
 22        "mean": 0.08337884861975908
 23      },
 24      {
 25        "cfg": {
 26          "lr": 0.006
 27        },
 28        "mean": 0.05978757236152887
 29      }
 30    ],
 31    "full": {
 32      "mean": 0.04506447445601225,
 33      "std": 0.01812126155730088,
 34      "per_seed": [
 35        0.06423857063055038,
 36        0.039343297481536865,
 37        0.05935552343726158,
 38        0.07621289789676666,
 39        0.028351932764053345,
 40        0.038264982402324677,
 41        0.0202757827937603,
 42        0.03447280824184418
 43      ],
 44      "n": 8
 45    }
 46  },
 47  "idea": {
 48    "mean": 0.05070888949558139,
 49    "std": 0.008596224148366128,
 50    "per_seed": [
 51      0.05224335566163063,
 52      0.03554942458868027,
 53      0.061948955059051514,
 54      0.04569929093122482,
 55      0.05899675190448761,
 56      0.05897253751754761,
 57      0.04991588369011879,
 58      0.042344916611909866
 59    ],
 60    "n": 8
 61  },
 62  "comparison": {
 63    "delta_mean": 0.0056444150395691395,
 64    "idea_wins": 3,
 65    "n_pairs": 8,
 66    "per_seed_diffs": [
 67      -0.011995214968919754,
 68      -0.003793872892856598,
 69      0.0025934316217899323,
 70      -0.03051360696554184,
 71      0.030644819140434265,
 72      0.02070755511522293,
 73      0.02964010089635849,
 74      0.007872108370065689
 75    ],
 76    "p_value": 0.49005,
 77    "mde": 0.01773643806217156,
 78    "mde_rel_pct": 39.35791613299345,
 79    "verdict": "no significant win",
 80    "system_worked": false
 81  },
 82  "mechanism_signature": {
 83    "predicted_budget": 0.537269855031944,
 84    "observed_budget_at_exit": 0.0,
 85    "observed_crossing_epoch": null,
 86    "observed_exit_epoch": null,
 87    "post_crossing_epochs": null,
 88    "relative_budget_error": Infinity,
 89    "confirmed": false
 90  },
 91  "idea_sweep": [
 92    {
 93      "cfg": {
 94        "lr": 0.001,
 95        "budget_scale": 1.0
 96      },
 97      "result": {
 98        "mean": 0.38905996084213257,
 99        "std": 0.06823595036576088,
100        "per_seed": [
101          0.4037257730960846,
102          0.27306991815567017,
103          0.47476810216903687,
104          0.4277902841567993,
105          0.4180760383605957,
106          0.460550993680954,
107          0.35112857818603516,
108          0.30336999893188477
109        ],
110        "n": 8
111      }
112    },
113    {
114      "cfg": {
115        "lr": 0.003,
116        "budget_scale": 1.0
117      },
118      "result": {
119        "mean": 0.07632979797199368,
120        "std": 0.02907599907859665,
121        "per_seed": [
122          0.07722286134958267,
123          0.03422033414244652,
124          0.11100310832262039,
125          0.061097800731658936,
126          0.10502404719591141,
127          0.11226481199264526,
128          0.0710601806640625,
129          0.03874523937702179
130        ],
131        "n": 8
132      }
133    },
134    {
135      "cfg": {
136        "lr": 0.006,
137        "budget_scale": 0.7
138      },
139      "result": {
140        "mean": 0.05070888949558139,
141        "std": 0.008596224148366128,
142        "per_seed": [
143          0.05224335566163063,
144          0.03554942458868027,
145          0.061948955059051514,
146          0.04569929093122482,
147          0.05899675190448761,
148          0.05897253751754761,
149          0.04991588369011879,
150          0.042344916611909866
151        ],
152        "n": 8
153      }
154    },
155    {
156      "cfg": {
157        "lr": 0.006,
158        "budget_scale": 1.0
159      },
160      "result": {
161        "mean": 0.05070888949558139,
162        "std": 0.008596224148366128,
163        "per_seed": [
164          0.05224335566163063,
165          0.03554942458868027,
166          0.061948955059051514,
167          0.04569929093122482,
168          0.05899675190448761,
169          0.05897253751754761,
170          0.04991588369011879,
171          0.042344916611909866
172        ],
173        "n": 8
174      }
175    },
176    {
177      "cfg": {
178        "lr": 0.006,
179        "budget_scale": 1.3
180      },
181      "result": {
182        "mean": 0.05070888949558139,
183        "std": 0.008596224148366128,
184        "per_seed": [
185          0.05224335566163063,
186          0.03554942458868027,
187          0.061948955059051514,
188          0.04569929093122482,
189          0.05899675190448761,
190          0.05897253751754761,
191          0.04991588369011879,
192          0.042344916611909866
193        ],
194        "n": 8
195      }
196    }
197  ],
198  "track_justification": "Dynamics track directly matches stability/control and bifurcation monitoring; shared rnn_small and task MSE.",
199  "budget": {
200    "n_train": 400,
201    "n_test": 200,
202    "epochs": 6,
203    "batch": 128,
204    "lr_union": [
205      0.001,
206      0.003,
207      0.006
208    ],
209    "paired_seeds": [
210      0,
211      1,
212      2,
213      3,
214      4,
215      5,
216      6,
217      7
218    ]
219  },
220  "runtime_sec": 64.41141510009766
221}