{% extends "base.html" %} {% block title %}Embeddings · {{ profile.name }}{% endblock %} {% block content %}

Embeddings benchmark · {{ profile.name }}

Benchmarks a batched-embedding request's latency and throughput (Lemonade's llamacpp GGUF / FastFlowLM NPU recipes, /v1/embeddings). Deliberately separate from the model/router leaderboard -- embeddings throughput isn't comparable to LLM token throughput. This runs synchronously.

{% if capabilities.error %}

Couldn't reach this instance ({{ capabilities.error }}) -- engine and backend fall back to free text.

{% endif %}
{% if embeddings_models %}
Lemonade doesn't flag which pulled models are embeddings-labeled -- pick one you know supports it. It errors clearly at run time if not.
{% else %}
No llamacpp/FastFlowLM models pulled yet on this profile -- pull one first, then reload this page.
{% endif %}
{% if embeddings_backends %} {% else %}
No backend detected -- install it on the profile page first, or enter it manually.
{% endif %}
{% if default_os %}
{{ default_os }} (fixed by this profile)
{% else %}
Couldn't detect -- see profile page to fix discovery.
{% endif %}
Each non-blank line becomes one item in the batch -- batch size materially affects latency.

Embeddings results for {{ profile.name }}

{% if not results %}

No embeddings results yet.

{% else %}
{% for r in results %} {% endfor %}
TimestampModelLatency (ms)Throughput (/sec)Batch sizeDimValidNotes
{{ r.timestamp }} {{ r.model.name }} {{ "%.1f"|format(r.metrics.latency_ms) }} {{ "%.2f"|format(r.metrics.embeddings_per_sec) }} {{ r.metrics.batch_size }} {{ r.metrics.embedding_dim or "–" }} {{ "valid" if r.validity.valid else "invalid" }} {{ r.validity.notes or "" }}
{% endif %}
{% endblock %}