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
122
123
124
125
126
127
128
129
|
{
"schema_version": 1,
"protocols": {
"burstccn_cifar10_symplastic_paper_seed0": {
"method": "BurstCCN",
"dataset": "CIFAR-10",
"seed": 0,
"expected_epochs": 400,
"repository_url": "https://github.com/neuralml/BurstCCN-journal",
"source_revision": "9170e110b4d18729dc5bc5d517fc52eefcd04d69",
"required_dataset_sha256": [
"6d958be074577803d12ecdefd02955f39262c83c16fe9348329d7fe0b5c001ce"
],
"expected_gpu": {
"physical_index": "5",
"name": "NVIDIA GeForce GTX 1080"
},
"paper_url": "https://www.biorxiv.org/content/10.64898/2026.06.16.732595v1",
"published_accuracy_percent_mean": 82.97,
"published_accuracy_percent_sd": 0.21,
"selection_semantics": {
"primary": "final epoch test accuracy",
"validation_selected": "test accuracy at the epoch with minimum validation top-1 error",
"test_selected": "maximum test accuracy over all epochs; reported separately"
},
"required_config": {
"model.loss_type": "mse",
"model.hidden_activation_function": "relu_rms",
"model.output_activation_function": "softmax",
"model.W.optimiser.lr": 0.002,
"model.W.optimiser.momentum": 0.9,
"model.W.optimiser.weight_decay": 0.0001,
"model.Y.mode": "sym_W",
"model.Y.grad_type": "kp",
"model.Y.optimiser.lr": 0.001,
"model.Q.mode": "sym_Y",
"model.Q.grad_type": "kp",
"model.Q.optimiser.lr": 0.001,
"dataset.dataset_name": "cifar10",
"dataset.use_validation": true,
"training.seed": 0,
"training.n_epochs": 400,
"training.batch_size": 32,
"training.label_smoothing": 0.05
},
"command": [
"python",
"train.py",
"--config-name=cifar10_burstccn_kp",
"dataset.root=/home/yurenh2/sdrn/data",
"training.seed=0",
"model.hidden_activation_function=relu_rms",
"model.output_activation_function=softmax",
"model.W.optimiser.lr=0.002",
"model.Y.mode=sym_W",
"model.Q.mode=sym_Y",
"hydra.run.dir=/scratch/yurenh2/burstccn-native/cifar10_symplastic_paper_seed0"
]
},
"dualprop_cifar10_vgg16_a0_b0p1_seed1988": {
"method": "Dual Propagation",
"dataset": "CIFAR-10",
"seed": 1988,
"expected_epochs": 130,
"repository_url": "https://github.com/Rasmuskh/dualprop_icml_2024",
"source_revision": "7b2595b34421e1483a721dbfdeff8cdabda3a1ff",
"required_dataset_sha256": [
"6d958be074577803d12ecdefd02955f39262c83c16fe9348329d7fe0b5c001ce",
"0b97b0a786c64f4cd0996cace20e81e189669aba9f4b4e517eb529f9c7105cec"
],
"expected_gpu": {
"physical_index": "7",
"name": "NVIDIA GeForce GTX 1080"
},
"paper_url": "https://arxiv.org/abs/2402.08573",
"published_accuracy_percent_mean": 92.41,
"published_accuracy_percent_sd": 0.07,
"selection_semantics": {
"primary": "single test evaluation from the best-validation checkpoint",
"validation_selected": "same as primary",
"test_selected": "not available; test is evaluated only once"
},
"resolved_config": {
"model": "VGG16",
"alpha": 0.0,
"beta": 0.1,
"learning_algorithm": "dualprop-lagr-ff",
"inference_sequence": "fwK",
"inference_passes_nudged": 16,
"batch_size": 100,
"learning_rate": 0.025,
"warmup_epochs": 10,
"decay_epochs": 120,
"momentum": 0.9,
"weight_decay": 0.0005,
"percent_train": 90,
"percent_val": 10
},
"command": [
"python",
"train.py",
"--model", "VGG16",
"--dataset", "cifar10",
"--num-epochs", "130",
"--batch-size", "100",
"--alpha", "0.0",
"--beta", "0.1",
"--loss", "sce",
"--inference-sequence", "fwK",
"--inference-passes-nudged", "16",
"--learning-rate", "0.025",
"--learning-rate-final", "0.000002",
"--warmup-learning-rate", "0.001",
"--learning-algorithm", "dualprop-lagr-ff",
"--activation", "relu",
"--decay-epochs", "120",
"--warmup-epochs", "10",
"--momentum", "0.9",
"--weight-decay", "5e-4",
"--dtype", "float32",
"--param-dtype", "float32",
"--percent-train", "90",
"--percent-val", "10",
"--seeds", "1988",
"--experiment-name", "sdil-dp-full-a0-b0p1-seed1988-v1"
]
}
}
}
|
