{% for dependency in bootstrap_js_dependencies %} {% endfor %}

{{title}}

{{subtitle}}

Model Version {{model_version_blessed}}Blessed !
{% if current_blessed_run_id is not none %} Last blessed run : run_id {{current_blessed_run_id}} ({{current_blessed_run_finished}}) here {% endif %} {% if local_serve_status is not none %}

Model Version {{local_serve_status}} Infra Validation !

{% endif %} {{local_serve_test_failed}}
{{records_count}} records : {{features_desc_table}}
This heatmap provides an overview of the relationships between features in the dataset and with the target variable. It does so by combining three different statistical metrics :
  • Correlation (Pearson's r) : Measures the linear relationship between pairs of numerical features. Values range from -1 to 1, where -1 indicates a perfect negative linear relationship, 0 indicates no linear relationship, and 1 indicates a perfect positive linear relationship.
  • Cramér's V : Measures the strength of association between pairs of categorical features. Values range from 0 to 1, where 0 indicates no association and 1 indicates a perfect association.
  • η² (Eta Squared) : Measures the proportion of variance in a numerical feature explained by a categorical feature. Values range from 0 to 1, where 0 indicates no explanatory power and 1 indicates perfect explanation.
section top ↑

Features bucketization :

{{buckets_table}}

Hyperparameters set :

{{hyperparameters_table}}

Performance metrics :

Overall

{{perf_metrics_table}} {% if styled_sliced_perf_metrics_table is not none %}

Per slice of {{slice_feature_name}}

{{styled_sliced_perf_metrics_table}} {% endif %}

Hyperparameters search space :

{% if hp_grid_table is not none %} {{hp_grid_table}} {{cv_folds}} cross-validation folds. {% else %} "
None
" {% endif %}
{% if wandb_project_ui_url is not none %} View more logs from these training jobs here on WandB i On the Weights&Biases workspace,
  • filter by mf_id = {{wandb_filter_run_id}}
  • group by "mf" (prefix for Metaflow task name).
. {% if "" != wandb_need_sync_dir %} Note that the herein flow run was not live-synced to WandB. Execute the following WandB CLI command locally to sync it so log-data shows there for this flow run :

wandb sync {{wandb_need_sync_dir}}

{% endif %} {% else %} Note that WandB logging was explicitely disabled for this flow run. {% endif %}
{% if hp_grid_table is not none %}

Hyperparameter tuning - Parallel Coordinates plot {{hp_perfs_curve}}
{{hp_perfs_table}}
{% endif %}
section top ↑

Training dataset :

# dataset size
{{task_obj_python_cmd}}['data_file'].size

# dataset object
{{task_obj_python_cmd}}['data_file'].data


Trained ML Model version :

{{task_obj_python_cmd}}['model'].data


Preprocessing Artifacts :

# num_features buckets edges
{{task_obj_python_cmd}}['buckets'].data

# fitted ordinal encoder
{{task_obj_python_cmd}}['encoder'].data.__dict__



section top ↑
{{dag_svg}}
section top ↑