{% from "_now1_macros.html" import heat_table with context %} {% macro comparisons(comparison, f) %} {% set bm = comparison.by_model %} {% if bm.columns %}
Runs by harness and model

a row per mode that ran, a column per harness and model with what it cost in all, each cell the mean cost of one run over its n runs, shaded and marked within the row the way every table here is

{{ heat_table(bm, bm.columns, bm.rows.keys() | list, bm.thin, bm.heads) }}
{% endif %} {% set tiers = comparison.tiers %}
By difficulty and model
{% if tiers.models %}

the five figures garden metrics reports, each cell with its n. Along a row the ground runs from green at the best value to red at the worst, and mark the two ends; a cell under {{ tiers.thin }} samples is faint and marked ~, and is never an end

{% for m in tiers.metrics %}{{ heat_table(m, tiers.models, ("easy", "medium", "hard"), tiers.thin) }}{% endfor %}
{% else %}

No model did work in this window.

{% endif %}
{% endmacro %}