{% extends "base.html" %} {% block title %}Run History - lm-eval-ledger{% endblock %} {% block content %}

Run History

ledger file: {{ db_file_bytes|fmtbytes }}
{% for r in runs %}
run {{ r.run_id }}{% if run_backend.get(r.run_id) %} · {{ run_backend[r.run_id] }}{% endif %} {{ r.started_at|fmtdate }} · {{ run_bytes.get(r.run_id, 0)|fmtbytes }} of samples
{% if r.source_yaml or r.config_yaml %}
config view download yaml
{{ r.source_yaml or r.config_yaml }}
{% endif %} {% set benches = by_run.get(r.run_id, []) %} {% if benches %} {% for b in benches %} {% set pending = b.accuracy is none and not b.error %} {% endfor %}
modeltask (few-shot)acc nno answersizetimetok/s
{{ b.model_tag }} {{ b.task }}{% if b.fewshot_k is not none %}({{ b.fewshot_k }}){% endif %} {% if b.verified_accuracy is not none %} {{ "%.3f"|format(b.verified_accuracy) }} {% else %}{{ "%.3f"|format(b.accuracy) if b.accuracy is not none else "…" }}{% endif %}{{ b.total_examples if b.total_examples is not none else ("…" if pending else "-") }} {{ b.no_answer_count if b.no_answer_count is not none else ("…" if pending else "-") }} {% set nb = bench_bytes.get(b.benchmark_id) %}{{ nb|fmtbytes if nb is not none else ("…" if pending else "-") }} {% if b.error %}ERROR{% elif b.duration_seconds is not none %}{{ b.duration_seconds|fmtdur }}{% elif pending and elapsed.get(b.benchmark_id) is not none %}{{ elapsed[b.benchmark_id]|fmtdur }} …{% else %}{{ "…" if pending else "-" }}{% endif %} {{ "%.0f"|format(b.gen_tokens / b.gen_seconds) if b.gen_tokens and b.gen_seconds else ("…" if pending else "-") }}
{% else %}

no benchmarks recorded

{% endif %}
{% else %}

The ledger is empty.

{% endfor %} {% endblock %}