{% extends "base.html" %} {% block content %}

Results

Evaluation metrics and feature importance for your trained models

{% if not runs %}
No results yet.
  1. Go to Run and check pipeline status.
  2. Approve the review_data gate and click Run Pipeline.
  3. Once training completes, results appear here.
{% else %} {# ── Run selector ── #}
Run: {% for r in runs %} {{ r.name }} {% endfor %} {% if runs | length > 1 %} Compare: {% for r in runs %}{% if r.name != selected_run %} vs {{ r.name }} {% endif %}{% endfor %} {% endif %}
{% if eval_report %} {# ── Metric cards ── #}
{% for k, v in eval_report.metrics.items() %}
{{ "%.4f" | format(v) }}
{{ k }}
{% if compare_report and k in compare_report.metrics %} {% set delta = compare_report.metrics[k] - v %}
{{ "%+.4f" | format(delta) }} vs {{ compare_run }}
{% endif %}
{% endfor %}
{# ── Feature importance ── #} {% if eval_report.feature_importance %}
Feature Importance (SHAP)
{% set max_fi = eval_report.feature_importance.values() | max %} {% for feat, score in eval_report.feature_importance.items() %}
{{ feat }}
{{ "%.3f" | format(score) }}
{% endfor %}

Higher = stronger influence. Low-importance columns can be dropped.

{% endif %} {# ── Next steps ── #}
What to do next

Happy? Go to Run, approve review_results, then click Run Pipeline to export.

Want to improve? Run aimodelground tune --trials 50 then re-run.

Compare runs: click another run button above, then append ?compare=run_001.

{# ── Compare table ── #} {% if compare_report %}
{{ selected_run }} vs {{ compare_run }}
{% for k in all_metrics %} {% set va = eval_report.metrics.get(k) %} {% set vb = compare_report.metrics.get(k) %} {% endfor %}
Metric {{ selected_run }} {{ compare_run }} Delta
{{ k }} {% if va is not none %}{{ "%.4f" | format(va) }}{% else %}—{% endif %} {% if vb is not none %}{{ "%.4f" | format(vb) }}{% else %}—{% endif %} {% if va is not none and vb is not none %} {% set delta = vb - va %} {{ "%+.4f" | format(delta) }} {% else %}—{% endif %}
{% endif %} {# ── Plotly chart ── #}
{% if eval_report.metrics %} {% endif %}
{% else %}
No eval report for {{ selected_run }} yet.
{% endif %} {% endif %} {% endblock %}