orbitalo/homelab-brain
1
0
Fork 0
Code Issues 37 Pull requests Projects Releases Packages Wiki Activity Actions

fix(llm): kein Fallback auf Textmodell bei Vision-Anfragen

Browse source 
Wenn qwen3-vl:32b timeout hat, wurde bisher auf qwen2.5:14b
(Textmodell) zurueckgefallen. Das kann keine Bilder sehen und
halluziniert stattdessen. Jetzt: allow_fallback=False fuer Vision
und klare Fehlermeldung bei Timeout.
This commit is contained in:
Homelab Cursor 2026-03-25 19:46:35 +01:00
parent ac0f6fda2e
commit 92898a33e3
1 changed files with 9 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

11
homelab-ai-bot/llm.py
View file

@ -525,7 +525,8 @@ def ask_with_image(image_base64: str, caption: str, tool_handlers: dict, session
try: try:
for _round in range(MAX_TOOL_ROUNDS): for _round in range(MAX_TOOL_ROUNDS):
data = _call_api(messages, api_key, use_tools=True, data = _call_api(messages, api_key, use_tools=True,
model=MODEL_VISION, max_tokens=4000) model=MODEL_VISION, max_tokens=4000,
allow_fallback=False)
choice = data["choices"][0] choice = data["choices"][0]
msg = choice["message"] msg = choice["message"]
@ -561,8 +562,14 @@ def ask_with_image(image_base64: str, caption: str, tool_handlers: dict, session
}) })
data = _call_api(messages, api_key, use_tools=False, data = _call_api(messages, api_key, use_tools=False,
model=MODEL_VISION, max_tokens=4000) model=MODEL_VISION, max_tokens=4000,
allow_fallback=False)
return data["choices"][0]["message"]["content"] return data["choices"][0]["message"]["content"]
except requests.exceptions.ReadTimeout:
return (
"Das Vision-Modell antwortet nicht (Timeout). "
"Bitte in 1-2 Min erneut versuchen — das Modell wird gerade geladen."
)
except Exception as e: except Exception as e:
return f"Vision-LLM-Fehler: {e}" return f"Vision-LLM-Fehler: {e}"