{% extends "base.html" %} {% block title %}Audio generation · {{ profile.name }}{% endblock %} {% block content %}
Benchmarks a music/sound-effect generation model's real-time-factor (Lemonade's acestep/thinksound
recipes, /v1/audio/generations -- not speech; see Text-to-speech for that).
Deliberately separate from the model/router leaderboard and from TTS results -- generating music is
a different task with a different compute profile. This runs synchronously.
Couldn't reach this instance ({{ capabilities.error }}) -- engine and backend fall back to free text.
{% endif %}No audio-generation results yet.
{% else %}| Timestamp | Model | Generation (ms) | Audio (s) | RTF | Valid | Notes |
|---|---|---|---|---|---|---|
| {{ r.timestamp }} | {{ r.model.name }} | {{ "%.1f"|format(r.metrics.generation_time_ms) }} | {{ "%.2f"|format(r.metrics.audio_duration_s) }} | {{ "%.2f"|format(r.metrics.real_time_factor) }}x | {{ "valid" if r.validity.valid else "invalid" }} | {{ r.validity.notes or "" }} |