alpha new hybridretriver line
This commit is contained in:
team2
@@ -22,7 +22,10 @@
|
||||
|
||||
<div class="row g-4">
|
||||
|
||||
<!-- Modell -->
|
||||
<!-- ============================== -->
|
||||
<!-- Modell-Basis -->
|
||||
<!-- ============================== -->
|
||||
|
||||
<div class="col-md-6">
|
||||
<label class="form-label">Modellname</label>
|
||||
<input type="text"
|
||||
@@ -31,11 +34,10 @@
|
||||
placeholder="z. B. qwen3:latest"
|
||||
required>
|
||||
<div class="form-text text-secondary">
|
||||
Exakter Modellname wie im KI-Endpunkt konfiguriert (z. B. Ollama oder API).
|
||||
Exakter Modellname wie im KI-Endpunkt konfiguriert.
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<!-- Stream -->
|
||||
<div class="col-md-6 d-flex align-items-center">
|
||||
<div class="form-check form-switch mt-4">
|
||||
<input class="form-check-input"
|
||||
@@ -48,11 +50,22 @@
|
||||
</label>
|
||||
</div>
|
||||
<div class="form-text text-secondary ms-3">
|
||||
Aktiviert Token-Streaming im Chat (empfohlen für bessere UX).
|
||||
Token-Streaming im Chat (empfohlen).
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<!-- Temperature -->
|
||||
</div>
|
||||
|
||||
<hr class="border-secondary my-4">
|
||||
|
||||
<!-- ============================== -->
|
||||
<!-- GENERATION -->
|
||||
<!-- ============================== -->
|
||||
|
||||
<h5 class="text-info mb-3">Generation (LLM Sampling)</h5>
|
||||
|
||||
<div class="row g-4">
|
||||
|
||||
<div class="col-md-4">
|
||||
<label class="form-label">Temperature</label>
|
||||
<input type="number"
|
||||
@@ -64,29 +77,24 @@
|
||||
class="form-control bg-dark text-light border-secondary"
|
||||
required>
|
||||
<div class="form-text text-secondary">
|
||||
Steuert die Kreativität der Antworten.
|
||||
Niedrige Werte (0.2–0.4) erzeugen stabile, sachliche Ergebnisse – empfohlen für RAG-Systeme.
|
||||
Höhere Werte führen zu freieren, weniger deterministischen Antworten.
|
||||
Kreativität des Modells. Für RAG 0.2–0.4 empfohlen.
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<!-- Top K -->
|
||||
<div class="col-md-4">
|
||||
<label class="form-label">Top K</label>
|
||||
<input type="number"
|
||||
min="1"
|
||||
max="200"
|
||||
name="top_k"
|
||||
value="40"
|
||||
class="form-control bg-dark text-light border-secondary"
|
||||
required>
|
||||
<div class="form-text text-secondary">
|
||||
Begrenzt die Anzahl der wahrscheinlichsten Token, aus denen das Modell auswählt.
|
||||
Niedrigere Werte = konservativer, höhere Werte = flexibler.
|
||||
20–50 ist für Wissenssysteme üblich.
|
||||
Begrenzt die Auswahl wahrscheinlicher Token.
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<!-- Top P -->
|
||||
<div class="col-md-4">
|
||||
<label class="form-label">Top P</label>
|
||||
<input type="number"
|
||||
@@ -98,13 +106,10 @@
|
||||
class="form-control bg-dark text-light border-secondary"
|
||||
required>
|
||||
<div class="form-text text-secondary">
|
||||
Nucleus Sampling: Das Modell berücksichtigt nur Token,
|
||||
deren kumulative Wahrscheinlichkeit innerhalb dieses Werts liegt.
|
||||
0.8–0.95 bietet eine gute Balance zwischen Stabilität und Natürlichkeit.
|
||||
Nucleus Sampling – Balance zwischen Stabilität und Natürlichkeit.
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<!-- Repeat Penalty -->
|
||||
<div class="col-md-6">
|
||||
<label class="form-label">Repeat Penalty</label>
|
||||
<input type="number"
|
||||
@@ -116,12 +121,10 @@
|
||||
class="form-control bg-dark text-light border-secondary"
|
||||
required>
|
||||
<div class="form-text text-secondary">
|
||||
Bestraft Wortwiederholungen. Werte leicht über 1.0 (z. B. 1.1–1.15)
|
||||
verhindern Schleifen und redundante Antworten.
|
||||
Verhindert Wiederholungen (1.1–1.15 empfohlen).
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<!-- Num Ctx -->
|
||||
<div class="col-md-6">
|
||||
<label class="form-label">Context Window (num_ctx)</label>
|
||||
<input type="number"
|
||||
@@ -132,9 +135,50 @@
|
||||
class="form-control bg-dark text-light border-secondary"
|
||||
required>
|
||||
<div class="form-text text-secondary">
|
||||
Maximale Kontextlänge in Tokens (Systemprompt + Benutzerfrage + Retrieval-Chunks).
|
||||
Muss vom Modell unterstützt werden.
|
||||
Höhere Werte ermöglichen größere Wissenskontexte, erhöhen jedoch Speicher- und Rechenbedarf.
|
||||
Maximale Kontextlänge (System + Frage + Retrieval).
|
||||
</div>
|
||||
</div>
|
||||
|
||||
</div>
|
||||
|
||||
<hr class="border-secondary my-4">
|
||||
|
||||
<!-- ============================== -->
|
||||
<!-- RETRIEVAL -->
|
||||
<!-- ============================== -->
|
||||
|
||||
<h5 class="text-warning mb-3">Retrieval (Wissensabruf)</h5>
|
||||
|
||||
<div class="row g-4">
|
||||
|
||||
<div class="col-md-6">
|
||||
<label class="form-label">Max Chunks</label>
|
||||
<input type="number"
|
||||
min="1"
|
||||
max="200"
|
||||
name="retrieval_max_chunks"
|
||||
value="25"
|
||||
class="form-control bg-dark text-light border-secondary"
|
||||
required>
|
||||
<div class="form-text text-secondary">
|
||||
Maximale Anzahl an Wissens-Chunks,
|
||||
die dem Modell übergeben werden.
|
||||
20–40 ist für die meisten Systeme optimal.
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<div class="col-md-6">
|
||||
<label class="form-label">Vector Top K</label>
|
||||
<input type="number"
|
||||
min="1"
|
||||
max="200"
|
||||
name="retrieval_vector_top_k"
|
||||
value="25"
|
||||
class="form-control bg-dark text-light border-secondary"
|
||||
required>
|
||||
<div class="form-text text-secondary">
|
||||
Anzahl der Vektor-Treffer vor Filterung.
|
||||
Höhere Werte erhöhen Recall, können aber Rauschen verstärken.
|
||||
</div>
|
||||
</div>
|
||||
|
||||
@@ -157,4 +201,4 @@
|
||||
Hinweis: Neue Konfigurationen werden zunächst inaktiv gespeichert und
|
||||
müssen separat aktiviert werden. Pro Modell kann nur eine Version aktiv sein.
|
||||
</div>
|
||||
{% endblock %}
|
||||
{% endblock %}
|
||||
@@ -24,11 +24,8 @@
|
||||
<th>Modell</th>
|
||||
<th>Version</th>
|
||||
<th>Stream</th>
|
||||
<th>Temp</th>
|
||||
<th>Top K</th>
|
||||
<th>Top P</th>
|
||||
<th>Repeat</th>
|
||||
<th>Ctx</th>
|
||||
<th>Sampling</th>
|
||||
<th class="text-warning">Retrieval</th>
|
||||
<th>Status</th>
|
||||
<th class="text-end">Aktionen</th>
|
||||
</tr>
|
||||
@@ -52,27 +49,47 @@
|
||||
{% endif %}
|
||||
</td>
|
||||
|
||||
<td>{{ config.temperature }}</td>
|
||||
<td>{{ config.topK }}</td>
|
||||
<td>{{ config.topP }}</td>
|
||||
<td>{{ config.repeatPenalty }}</td>
|
||||
<td>{{ config.numCtx }}</td>
|
||||
{# ========================= #}
|
||||
{# Sampling Group #}
|
||||
{# ========================= #}
|
||||
<td>
|
||||
<div class="small">
|
||||
<div><strong>T:</strong> {{ config.temperature }}</div>
|
||||
<div><strong>K:</strong> {{ config.topK }}</div>
|
||||
<div><strong>P:</strong> {{ config.topP }}</div>
|
||||
<div><strong>R:</strong> {{ config.repeatPenalty }}</div>
|
||||
<div><strong>Ctx:</strong> {{ config.numCtx }}</div>
|
||||
</div>
|
||||
</td>
|
||||
|
||||
{# ========================= #}
|
||||
{# Retrieval Group #}
|
||||
{# ========================= #}
|
||||
<td class="text-warning">
|
||||
<div class="small">
|
||||
<div><strong>Chunks:</strong> {{ config.retrievalMaxChunks }}</div>
|
||||
<div><strong>VectorK:</strong> {{ config.retrievalVectorTopK }}</div>
|
||||
</div>
|
||||
</td>
|
||||
|
||||
<td>
|
||||
{% if config.active %}
|
||||
<span class="badge bg-success">Aktiv</span>
|
||||
{% else %}
|
||||
<span class="badge bg-dark border border-secondary">
|
||||
Inaktiv
|
||||
</span>
|
||||
Inaktiv
|
||||
</span>
|
||||
{% endif %}
|
||||
</td>
|
||||
|
||||
<td class="text-end">
|
||||
<a href="{{ path('admin_model_config_test_retrieval', {id: config.id}) }}"
|
||||
class="btn btn-sm btn-outline-warning me-1">
|
||||
Test Retrieval
|
||||
</a>
|
||||
|
||||
{% if not config.active and is_granted('ROLE_SUPER_ADMIN') %}
|
||||
|
||||
{# Aktivieren via POST + CSRF #}
|
||||
<form method="post"
|
||||
action="{{ path('admin_model_config_activate', {id: config.id}) }}"
|
||||
class="d-inline"
|
||||
@@ -82,12 +99,11 @@
|
||||
name="_token"
|
||||
value="{{ csrf_token('activate_model_config_' ~ config.id) }}">
|
||||
|
||||
<button class="btn btn-sm btn-outline-success me-2">
|
||||
<button class="btn btn-sm btn-outline-success me-1">
|
||||
Aktivieren
|
||||
</button>
|
||||
</form>
|
||||
|
||||
{# Löschen via POST + CSRF #}
|
||||
<form method="post"
|
||||
action="{{ path('admin_model_config_delete', {id: config.id}) }}"
|
||||
class="d-inline"
|
||||
@@ -105,13 +121,12 @@
|
||||
{% else %}
|
||||
<span class="text-secondary small">—</span>
|
||||
{% endif %}
|
||||
|
||||
</td>
|
||||
</tr>
|
||||
|
||||
{% else %}
|
||||
<tr>
|
||||
<td colspan="10" class="text-center text-secondary py-4">
|
||||
<td colspan="7" class="text-center text-secondary py-4">
|
||||
Keine Konfiguration vorhanden.
|
||||
</td>
|
||||
</tr>
|
||||
@@ -132,13 +147,11 @@
|
||||
|
||||
<div class="card bg-black border-info">
|
||||
<div class="card-body p-0">
|
||||
|
||||
<iframe
|
||||
src="/index.html?admin_test=1"
|
||||
class="w-100 border-0"
|
||||
style="height:75vh;"
|
||||
></iframe>
|
||||
|
||||
</div>
|
||||
</div>
|
||||
|
||||
@@ -146,4 +159,5 @@
|
||||
Der Agent läuft im isolierten Admin-Test-Modus.
|
||||
Keine Persistenz. Keine produktive Session.
|
||||
</div>
|
||||
{% endblock %}
|
||||
|
||||
{% endblock %}
|
||||
93
templates/admin/model_config/test_retrieval.html.twig
Normal file
93
templates/admin/model_config/test_retrieval.html.twig
Normal file
@@ -0,0 +1,93 @@
|
||||
{% extends 'admin/base.html.twig' %}
|
||||
|
||||
{% block title %}Retrieval-Test{% endblock %}
|
||||
|
||||
{% block body %}
|
||||
|
||||
<div class="d-flex justify-content-between align-items-center mb-4">
|
||||
<h1 class="h3">
|
||||
Retrieval-Test – {{ config.modelName }} (v{{ config.version }})
|
||||
</h1>
|
||||
|
||||
<a href="{{ path('admin_model_config_list') }}"
|
||||
class="btn btn-sm btn-outline-secondary">
|
||||
Zurück
|
||||
</a>
|
||||
</div>
|
||||
|
||||
{# ====================================================== #}
|
||||
{# Konfigurationsübersicht #}
|
||||
{# ====================================================== #}
|
||||
|
||||
<div class="card bg-black border-secondary text-light mb-4">
|
||||
<div class="card-body small">
|
||||
|
||||
<div class="row">
|
||||
<div class="col-md-6">
|
||||
<strong>Max Chunks:</strong>
|
||||
{{ config.retrievalMaxChunks }}
|
||||
</div>
|
||||
|
||||
<div class="col-md-6">
|
||||
<strong>Vector Top K:</strong>
|
||||
{{ config.retrievalVectorTopK }}
|
||||
</div>
|
||||
</div>
|
||||
|
||||
</div>
|
||||
</div>
|
||||
|
||||
{# ====================================================== #}
|
||||
{# Testformular #}
|
||||
{# ====================================================== #}
|
||||
|
||||
<div class="card bg-black border-secondary text-light mb-4">
|
||||
<div class="card-body">
|
||||
|
||||
<form method="post">
|
||||
|
||||
<div class="mb-3">
|
||||
<label class="form-label">Test-Prompt</label>
|
||||
<textarea name="prompt"
|
||||
rows="3"
|
||||
class="form-control bg-dark text-light border-secondary"
|
||||
required>{{ prompt }}</textarea>
|
||||
</div>
|
||||
|
||||
<button type="submit"
|
||||
class="btn btn-outline-warning">
|
||||
Retrieval testen
|
||||
</button>
|
||||
|
||||
</form>
|
||||
|
||||
</div>
|
||||
</div>
|
||||
|
||||
{# ====================================================== #}
|
||||
{# Ergebnisse #}
|
||||
{# ====================================================== #}
|
||||
|
||||
{% if results is not empty %}
|
||||
<div class="card bg-black border-secondary text-light">
|
||||
<div class="card-body">
|
||||
|
||||
<h5 class="text-warning mb-3">
|
||||
Gefundene Chunks ({{ results|length }})
|
||||
</h5>
|
||||
|
||||
<div style="max-height: 500px; overflow-y: auto;">
|
||||
|
||||
{% for chunk in results %}
|
||||
<div class="border border-secondary p-3 mb-3 small">
|
||||
{{ chunk }}
|
||||
</div>
|
||||
{% endfor %}
|
||||
|
||||
</div>
|
||||
|
||||
</div>
|
||||
</div>
|
||||
{% endif %}
|
||||
|
||||
{% endblock %}
|
||||
Reference in New Issue
Block a user