Dom/rpa_vision_v3
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(grounding): désactiver orchestrateur VRAM pendant exécution + qwen2.5vl:3b pour description
Some checks failed
security-audit / Bandit (scan statique) (push) Successful in 16s
Details
security-audit / pip-audit (CVE dépendances) (push) Successful in 10s
Details
security-audit / Scan secrets (grep) (push) Successful in 8s
Details
tests / Lint (ruff + black) (push) Successful in 13s
Details
tests / Tests unitaires (sans GPU) (push) Failing after 14s
Details
tests / Tests sécurité (critique) (push) Has been skipped
Details

Browse Source 
L'orchestrateur VRAM redémarrait Ollama en pleine exécution → timeout.
Désactivé pendant le workflow. L'orchestrateur reste disponible pour
bascule manuelle avant/après.

Description ancre via qwen2.5vl:3b (3 Go) au lieu de 7b — tient en VRAM
sans décharger CLIP ni RF-DETR.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
Dom
2026-04-21 10:16:27 +02:00
parent d1b556b6cd
commit 203e5cc6c1
1 changed files with 3 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
core/execution/input_handler.py
View File

@@ -286,10 +286,6 @@ Si tu vois un dialogue ou une popup, indique quel bouton cliquer.
Si l'écran est normal sans action nécessaire, réponds action="nothing". Si l'écran est normal sans action nécessaire, réponds action="nothing".
Réponds UNIQUEMENT le JSON, pas d'explication.""" Réponds UNIQUEMENT le JSON, pas d'explication."""
from core.cognition.vram_orchestrator import get_orchestrator
orch = get_orchestrator()
orch.ensure_reasoning_ready()
ollama_url = os.environ.get("OLLAMA_URL", "http://localhost:11434") ollama_url = os.environ.get("OLLAMA_URL", "http://localhost:11434")
model = os.environ.get("RPA_REASONING_MODEL", "qwen2.5vl:7b") model = os.environ.get("RPA_REASONING_MODEL", "qwen2.5vl:7b")
@@ -402,8 +398,9 @@ def _describe_anchor_image(anchor_image_base64: str) -> Optional[str]:
anchor_image_base64 = anchor_image_base64.split(',', 1)[1] anchor_image_base64 = anchor_image_base64.split(',', 1)[1]
ollama_url = os.environ.get("OLLAMA_URL", "http://localhost:11434") ollama_url = os.environ.get("OLLAMA_URL", "http://localhost:11434")
model = os.environ.get("RPA_REASONING_MODEL", "qwen2.5vl:7b") model = "qwen2.5vl:3b"
logger.info(f"[Grounding] Description ancre via {model}...")
response = requests.post( response = requests.post(
f"{ollama_url}/api/generate", f"{ollama_url}/api/generate",
json={ json={
@@ -413,7 +410,7 @@ def _describe_anchor_image(anchor_image_base64: str) -> Optional[str]:
"stream": False, "stream": False,
"options": {"temperature": 0.1, "num_predict": 20} "options": {"temperature": 0.1, "num_predict": 20}
}, },
timeout=15 timeout=30
) )
if response.status_code == 200: if response.status_code == 200: