Bi-Maxwell Muon / verification.json
Unverified
1{
2 "predictions": {
3 "half_life": [
4 {
5 "beta": 0.6065306597126334,
6 "predicted_continuous_index": 0.3862943611198906,
7 "predicted_first_integer": 1,
8 "observed_first_integer": 1
9 },
10 {
11 "beta": 0.951229424500714,
12 "predicted_continuous_index": 12.862943611198906,
13 "predicted_first_integer": 13,
14 "observed_first_integer": 13
15 },
16 {
17 "beta": 0.8,
18 "predicted_continuous_index": 2.1062837195053903,
19 "predicted_first_integer": 3,
20 "observed_first_integer": 3
21 },
22 {
23 "beta": 0.9,
24 "predicted_continuous_index": 5.578813478960585,
25 "predicted_first_integer": 6,
26 "observed_first_integer": 6
27 },
28 {
29 "beta": 0.99,
30 "predicted_continuous_index": 67.96756393652842,
31 "predicted_first_integer": 68,
32 "observed_first_integer": 68
33 }
34 ],
35 "impulse_log_slope": [
36 {
37 "beta": 0.5,
38 "predicted_log_slope": -0.6931471805599453,
39 "observed_log_slope": -0.6931471805599451
40 },
41 {
42 "beta": 0.8,
43 "predicted_log_slope": -0.2231435513142097,
44 "observed_log_slope": -0.22314355131420968
45 },
46 {
47 "beta": 0.9,
48 "predicted_log_slope": -0.10536051565782628,
49 "observed_log_slope": -0.10536051565782625
50 },
51 {
52 "beta": 0.95,
53 "predicted_log_slope": -0.05129329438755058,
54 "observed_log_slope": -0.05129329438755058
55 },
56 {
57 "beta": 0.99,
58 "predicted_log_slope": -0.01005033585350145,
59 "observed_log_slope": -0.01005033585350145
60 }
61 ],
62 "frequency_gain": [
63 {
64 "omega": 0.1,
65 "predicted_gain": 0.7141921373403015,
66 "observed_gain": 0.6527129624755074,
67 "relative_error": 0.08608212223358638
68 },
69 {
70 "omega": 0.5,
71 "predicted_gain": 0.4074906029120214,
72 "observed_gain": 0.3979262294028173,
73 "relative_error": 0.023471396495661188
74 },
75 {
76 "omega": 1.5,
77 "predicted_gain": 0.19207777233082604,
78 "observed_gain": 0.19164946313229175,
79 "relative_error": 0.0022298738335875096
80 },
81 {
82 "omega": 2.5,
83 "predicted_gain": 0.14178606426188498,
84 "observed_gain": 0.14175263539042465,
85 "relative_error": 0.0002357697960963076
86 }
87 ]
88 },
89 "mini_experiment": {
90 "Muon": {
91 "final_loss": 0.0015966509849844191,
92 "loss_step_100": 1.532585040927958,
93 "mean_gradient_update_cosine": 0.5865445897775171,
94 "seconds": 0.1414241020102054
95 },
96 "BiMaxwell_(.90,.99,.5)": {
97 "final_loss": 0.004175016413905212,
98 "loss_step_100": 1.533368522066344,
99 "mean_gradient_update_cosine": 0.5501982748971043,
100 "seconds": 0.13702032099536154
101 },
102 "BiMaxwell_(.95,.995,.5)": {
103 "final_loss": 0.023674204360653456,
104 "loss_step_100": 1.537552271016059,
105 "mean_gradient_update_cosine": 0.5017544363226999,
106 "seconds": 0.15190634499595035
107 },
108 "AdamW": {
109 "final_loss": 0.0011776805452452907,
110 "loss_step_100": 0.0012634382348207475,
111 "mean_gradient_update_cosine": 0.344865794518648,
112 "seconds": 0.06803519600362051
113 }
114 }
115}