-
Notifications
You must be signed in to change notification settings - Fork 0
/
configuration.yaml
143 lines (117 loc) · 3.36 KB
/
configuration.yaml
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
130
131
132
133
134
135
136
137
138
139
140
141
142
143
# general
################################################################################
# experiment files directory
save_directory: experiments
# experiment naming prefix
experiment_prefix: transformer
################################################################################
# dataset
################################################################################
num_symbols: 3
#num_symbols: 100
#num_symbols: 1000
#min_frequency: 300 # ~ 7 symbols
#min_frequency: 200 # ~ 4000 symbols
#min_frequency: 100 # ~ 14000 symbols
#min_frequency: 24 # ~ 25000 symbols
#min_frequency: 10 # ~ 30500 symbols
# exclude genera from all pipeline datasets
excluded_genera:
# Cow (genus Bos)
#- Bos
# Horse, Donkey (genus Equus)
#- Equus
# salmon and trout (both in genera Salmo and Oncorhynchus)
#- Salmo
#- Oncorhynchus
# Zebra finch (genus Taeniopygia)
#- Taeniopygia
# Chinese hamster (genus Cricetulus)
#- Cricetulus
# Macaque (genus Macaca) (Rhesus macaque)
#- Macaca
# training, validation, test split
test_ratio: 0.2
validation_ratio: 0.2
#test_ratio: 0.15
#validation_ratio: 0.15
#test_ratio: 0.1
#validation_ratio: 0.1
#test_ratio: 0
#validation_ratio: 0.1
################################################################################
# features
################################################################################
sequence_length: 581 # ~ 581.36 : dataset sequences mean length
#sequence_length: 844 # ~ 581.36 + 0.5 * 526.05 : mean + 0.5 * standard_deviation
#sequence_length: 1107 # ~ 581.36 + 1 * 526.05 : mean + 1 * standard_deviation
#padding_side: left
padding_side: right
clade: True
#clade: False
################################################################################
# network architecture
################################################################################
#mlp_output_size: 32
mlp_output_size: 64
#mlp_output_size: 128
embedding_dimension: 8
#embedding_dimension: 16
#embedding_dimension: 32
#embedding_dimension: 64
num_heads: 2
#num_heads: 4
#num_heads: 8
transformer_depth: 1
#transformer_depth: 2
#transformer_depth: 3
activation_function: relu
#activation_function: gelu
#feedforward_connections: 64
#feedforward_connections: 128
feedforward_connections: 256
# L2 regularization
weight_decay: 0
#weight_decay: 1.0e-6
#weight_decay: 1.0e-5
# max norm for gradient clipping
clip_max_norm: 0
#clip_max_norm: 5
dropout_probability: 0
#dropout_probability: 0.3
#dropout_probability: 0.5
################################################################################
# training
################################################################################
# random number generator initialization seed
random_seed: 5
#random_seed: 7
#random_seed: 11
batch_size: 16
#batch_size: 32
#batch_size: 64
num_workers: 0
#num_workers: 1
#num_workers: 5
#num_workers: 17
learning_rate: 3.0e-4
#learning_rate: 1.0e-4
#learning_rate: 3.0e-5
# number of epochs without validation loss improvement before training stops
patience: 3
#patience: 11
#patience: 31
# minimum validation loss change to consider as improvement
loss_delta: 0
#loss_delta: 1.0e-6
# maximum number of training epochs
#max_epochs: 1
#max_epochs: 3
max_epochs: 10
#max_epochs: 100
#max_epochs: 1000
profiler: null
#profiler: simple
#profiler: pytorch
num_sample_predictions: 10
################################################################################