-
Notifications
You must be signed in to change notification settings - Fork 0
/
pipeline-old.yaml
146 lines (130 loc) · 4.38 KB
/
pipeline-old.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
144
145
146
meta:
# import_tasks_from: features.yaml
# Extract upstream dependencies from source code. If False, tasks
# must declare dependencies using the "upstream" key
extract_upstream: False
# Extract product from source code. If False, tasks must have a "product" key
extract_product: False
executor: serial
tasks:
- source: func.util.data_transform
name: data-transform
upstream: null
product: products/data/data.csv
params:
file_path: '{{root}}/{{DATA_FILEPATH}}'
- source: scripts/features.py
name: features
upstream: [data-transform]
product:
data: products/data/features.csv
nb: products/reports/features.ipynb
params:
features: '{{features}}'
target: '{{TARGET}}'
- source: scripts/features.py
name: features-alt
upstream: [ data-transform ]
product:
data: products/data/features_alt.csv
nb: products/reports/features_alt.ipynb
params:
features: '{{features}}'
target: '{{TARGET_ALT}}'
- source: func.util.split_train_test_op_alt
name: split-train-test-alt
upstream: [ features-alt ]
product:
train: products/data/train_alt.csv
test: products/data/test_alt.csv
train_odds: products/data/train_odds_alt.csv
test_odds: products/data/test_odds_alt.csv
params:
test_ratio: '{{TEST_RATIO}}'
- source: scripts/fit_pytorch_alt.py
name: fit-pytorch-alt
upstream: [ split-train-test-alt ]
product:
nb: products/reports/fit_pytorch_alt.ipynb
model_state_dict: products/models/pytorch_state_dict_alt.pt
model: products/models/pytorch_alt.pt
params:
target: '{{TARGET_ALT}}'
pytorch_conf: '{{pytorch_conf}}'
random_seed: '{{RANDOM_SEED}}'
validation_ratio: 0.2
odds_cols: ''
- source: scripts/eval_pytorch_alt.py
name: eval-pytorch-alt
upstream: [ split-train-test-alt, fit-pytorch-alt ]
product:
nb: products/reports/eval_pytorch_alt.ipynb
params:
target: '{{TARGET_ALT}}'
pytorch_conf: '{{pytorch_conf}}'
odds_cols: '{{odds_cols_alt}}'
bootstrap_repetitions: 3
kelly_fraction: 0.05
- source: func.util.split_train_test_op
name: split-train-test
upstream: [features]
product:
train: products/data/train.csv
test: products/data/test.csv
train_odds: products/data/train_odds.csv
test_odds: products/data/test_odds.csv
params:
test_ratio: '{{TEST_RATIO}}'
- source: scripts/fit_sklearn_automl.py
name: automl-sklearn
upstream: [split-train-test]
product:
nb: products/reports/fit-sklearn-automl.ipynb
model: products/models/sklearn-automl.pickle
params:
target: '{{TARGET}}'
random_seed: '{{RANDOM_SEED}}'
autosklearn_config: '{{autosklearn_config}}'
- source: scripts/fit_h2o_automl.py
name: automl-h2o
upstream: [split-train-test]
product:
nb: products/reports/fit_h2o_automl.ipynb
params:
target: '{{TARGET}}'
random_seed: '{{RANDOM_SEED}}'
factors: '{{factors}}'
h2oautoml_config: '{{h2oautoml_config}}'
models_path: '{{root}}/products/models/h2o'
- source: ntb/automl_evaluation.ipynb
name: automl-eval
upstream: [automl-sklearn, automl-h2o, split-train-test]
product:
nb: products/reports/automl_evaluation.ipynb
autosklearn_matrix: products/reports/autosklearn_matrix.html
params:
target: '{{TARGET}}'
- source: scripts/fit_pytorch.py
name: fit-pytorch
upstream: [split-train-test]
product:
nb: products/reports/fit_pytorch.ipynb
model_state_dict: products/models/pytorch_state_dict.pt
model: products/models/pytorch.pt
params:
target: '{{TARGET}}'
pytorch_conf: '{{pytorch_conf}}'
random_seed: '{{RANDOM_SEED}}'
validation_ratio: 0.2
odds_cols: ''
- source: scripts/eval_pytorch.py
name: eval-pytorch
upstream: [split-train-test, fit-pytorch]
product:
nb: products/reports/eval_pytorch.ipynb
params:
target: '{{TARGET}}'
pytorch_conf: '{{pytorch_conf}}'
odds_cols: '{{odds_cols}}'
bootstrap_repetitions: 3
kelly_fraction: 0.05