1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
|
{
"schema": "rain_ep_structured_bias_screen_v1",
"sdil": {
"revision": "a10a84066a6f7aa38d86b1a71ed371c9c6f22815"
},
"author": {
"repository": "https://github.com/rain-neuromorphics/energy-based-learning",
"revision": "6b253fd8a5d267535f58ab79992256ef10031ceb"
},
"protocol": {
"dataset": "FashionMNIST",
"network": "author ConvHopfieldEnergy28 32-64-10",
"algorithm": "positive equilibrium propagation",
"adapter": "layer",
"mode": "same_rms_noise",
"bias_ratio": 0.01,
"predictor_rate": 0.2,
"neutral_cadence": 1,
"layer_calibration_steps": 1,
"calibration_batches": 0,
"calibration_observations": 0,
"calibration_seconds": 4.76837158203125e-07,
"extra_equilibrium_phases_for_predictor": 0,
"predictor_neutral_source": "existing_first_EP_phase",
"bias_ratio_normalization": "experimenter_initial_clean_layer_state_difference_rms",
"bias_ratio_normalization_visible_to_predictor": false,
"epochs": 3,
"train_limit": 10000,
"test_limit": 2000,
"evaluation_split": "train_holdout",
"data_seed": 6100,
"batch_size": 128,
"training_iterations": 12,
"inference_iterations": 30,
"seed": 1989,
"device": "cuda",
"determinism": "author_default",
"autodiff_used_for_learning": false
},
"hardware": {
"torch_version": "2.3.1+cu118",
"torch_cuda_version": "11.8",
"cuda_visible_devices": "5",
"device_name": "NVIDIA GeForce GTX 1080"
},
"metrics": [
{
"epoch": 1,
"train_accuracy": 0.6017,
"train_cost": 0.30215022211074827,
"test_accuracy": 0.761,
"test_cost": 0.2138809404373169,
"finite": true,
"corrector": {
"step": 78,
"clean_state_difference_rms": 0.002920974688811938,
"bias_state_difference_rms": 6.663539274315504e-05,
"residual_state_difference_rms": 6.636989063542658e-05,
"bias_to_clean_state_difference_rms": 0.022812725149035092,
"residual_to_clean_state_difference_rms": 0.02272183011021623,
"neutral_observations": 0
},
"wall_seconds": 11.205537796020508
},
{
"epoch": 2,
"train_accuracy": 0.7882,
"train_cost": 0.188776060628891,
"test_accuracy": 0.8015,
"test_cost": 0.1777397050857544,
"finite": true,
"corrector": {
"step": 157,
"clean_state_difference_rms": 0.0032720191567634283,
"bias_state_difference_rms": 6.637821262985081e-05,
"residual_state_difference_rms": 6.59492970724865e-05,
"bias_to_clean_state_difference_rms": 0.020286621028071827,
"residual_to_clean_state_difference_rms": 0.020155535133761667,
"neutral_observations": 0
},
"wall_seconds": 21.835002899169922
},
{
"epoch": 3,
"train_accuracy": 0.822,
"train_cost": 0.16819184494018555,
"test_accuracy": 0.7965,
"test_cost": 0.18230166244506835,
"finite": true,
"corrector": {
"step": 236,
"clean_state_difference_rms": 0.0061238763125476565,
"bias_state_difference_rms": 6.725749249790725e-05,
"residual_state_difference_rms": 6.796068294087145e-05,
"bias_to_clean_state_difference_rms": 0.010982830002640399,
"residual_to_clean_state_difference_rms": 0.011097657671762877,
"neutral_observations": 0
},
"wall_seconds": 32.36355447769165
}
],
"epochs_completed": 3,
"final": {
"epoch": 3,
"train_accuracy": 0.822,
"train_cost": 0.16819184494018555,
"test_accuracy": 0.7965,
"test_cost": 0.18230166244506835,
"finite": true,
"corrector": {
"step": 236,
"clean_state_difference_rms": 0.0061238763125476565,
"bias_state_difference_rms": 6.725749249790725e-05,
"residual_state_difference_rms": 6.796068294087145e-05,
"bias_to_clean_state_difference_rms": 0.010982830002640399,
"residual_to_clean_state_difference_rms": 0.011097657671762877,
"neutral_observations": 0
},
"wall_seconds": 32.36355447769165
}
}
|
