Second-order SCAFFOLD bias compensation / bench_report.json
Failed on benchmark
1{
2 "bench_version": 1,
3 "track": "tabular",
4 "model": "mlp_tiny",
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": 0.0015
15 },
16 "mean": 6.50612199306488
17 },
18 {
19 "cfg": {
20 "lr": 0.003
21 },
22 "mean": 5.869555473327637
23 },
24 {
25 "cfg": {
26 "lr": 0.006
27 },
28 "mean": 4.962810158729553
29 }
30 ],
31 "full": {
32 "mean": 5.057340204715729,
33 "std": 0.5639876094204693,
34 "per_seed": [
35 5.718886852264404,
36 5.579864501953125,
37 4.7256927490234375,
38 3.826796531677246,
39 5.060815811157227,
40 5.323063373565674,
41 4.853731155395508,
42 5.369870662689209
43 ],
44 "n": 8
45 }
46 },
47 "idea": {
48 "mean": 4.5547012984752655,
49 "std": 0.8416642558828814,
50 "per_seed": [
51 5.184708118438721,
52 5.024646759033203,
53 4.222005367279053,
54 4.777006149291992,
55 5.235812664031982,
56 4.5309157371521,
57 4.969952583312988,
58 2.492563009262085
59 ],
60 "n": 8
61 },
62 "comparison": {
63 "delta_mean": -0.5026389062404633,
64 "idea_wins": 5,
65 "n_pairs": 8,
66 "per_seed_diffs": [
67 -0.5341787338256836,
68 -0.5552177429199219,
69 -0.5036873817443848,
70 0.9502096176147461,
71 0.17499685287475586,
72 -0.7921476364135742,
73 0.11622142791748047,
74 -2.877307653427124
75 ],
76 "p_value": 0.2491,
77 "mde": 0.9310786105933415,
78 "mde_rel_pct": 18.4104405261318,
79 "verdict": "no significant win",
80 "system_worked": false
81 },
82 "idea_sweep": [
83 {
84 "cfg": {
85 "lr": 0.0015,
86 "delta": 0.02,
87 "ema": 0.15
88 },
89 "result": {
90 "mean": 6.439422070980072,
91 "std": 0.22122740568514504,
92 "per_seed": [
93 6.449733257293701,
94 6.59368371963501,
95 6.736842155456543,
96 6.062600612640381,
97 6.195968151092529,
98 6.431543350219727,
99 6.701443672180176,
100 6.34356164932251
101 ],
102 "n": 8
103 }
104 },
105 {
106 "cfg": {
107 "lr": 0.003,
108 "delta": 0.02,
109 "ema": 0.15
110 },
111 "result": {
112 "mean": 5.547695994377136,
113 "std": 0.3737932778729164,
114 "per_seed": [
115 5.546213150024414,
116 6.377541542053223,
117 5.3842549324035645,
118 5.396069526672363,
119 5.144563674926758,
120 5.520054817199707,
121 5.180293083190918,
122 5.832577228546143
123 ],
124 "n": 8
125 }
126 },
127 {
128 "cfg": {
129 "lr": 0.006,
130 "delta": 0.01,
131 "ema": 0.1
132 },
133 "result": {
134 "mean": 4.5547012984752655,
135 "std": 0.8416642558828814,
136 "per_seed": [
137 5.184708118438721,
138 5.024646759033203,
139 4.222005367279053,
140 4.777006149291992,
141 5.235812664031982,
142 4.5309157371521,
143 4.969952583312988,
144 2.492563009262085
145 ],
146 "n": 8
147 }
148 }
149 ],
150 "selected_cfg": {
151 "lr": 0.006,
152 "delta": 0.01,
153 "ema": 0.1
154 },
155 "custom_track": null,
156 "mechanism_signature": {
157 "f2_hat": 0.12070295963746307,
158 "f3_hat": -1.2520681114060102,
159 "sigma2_hat": 6.631455711858227e-06,
160 "predicted_displacement": -3.4351747621157183e-11,
161 "observed_displacement": 5.153735449760493e-08,
162 "n_obs": 480,
163 "confirmed": false
164 }
165}