mle-deployment-example/config.yaml at main · ColtAllen/mle-deployment-example · GitHub

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
gcp:
  project_id: colt-allen-code-challenge
  region: europe-west3
  dataset_id: bq_code_challenge
  feature_table_id: mart_features
  target_table_id: mart_targets

training:
  target_classes:
    0: neither
    1: group_1
    2: group_2
  categorical_features:
  - c_5
  - c_6
  - c_7
  - c_8
  - c_24
  drop_columns:
  - ID
  - LOAD_TIMESTAMP
  - row_modified
  - modification_type
  k_folds_cv: 5
  classification_avg: weighted avg
  confusion_normalize: all
  coerce_params:
  - combinations_ctr
  - simple_ctr
  - pool_metainfo_options
  - class_names
  - class_weights
  metric_payoff:
    correct_target: 100
    wrong_target: 50
    correct_no_target: 0
    wrong_no_target: -50
  metadata_table_id: dim_feature_metadata

catboost:
  loss_function: MultiClass
  eval_metric: Accuracy
  early_stopping_rounds: 20
  langevin: true
  auto_class_weights:
  random_seed:
  verbose: false
  allow_writing_files: false
  task_type: CPU
  thread_count:

vertex_ai:
  experiment_name: code-challenge-catboost
  experiment_description: prod run of catboost model
  staging_bucket: code-challenge-bucket
  model_uri_path: models/latest_uri.txt
  model_gcs_path: catboost
  plots_gcs_path: plots

data_etl:
  source_path: raw.parquet
  raw_table_id: bq_raw
  n_samples: 1000
  drift_probability: 0.3
  gcs_bucket: code-challenge-bucket
  seed:
  if_exists: append

serving:
  prediction_type: Class
  max_batch_size: 100