Dom/t2a
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: embeddings sur GPU (CUDA) pour l'indexation et la recherche RAG

Browse Source 
Détection automatique GPU/CPU avec fallback. Index FAISS reconstruit
en 1min (GPU) au lieu de 16min (CPU).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
dom
2026-02-11 23:42:46 +01:00
parent b38f87ac7a
commit 931b6c5d1c
2 changed files with 8 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/medical/rag_search.py
View File

@@ -29,7 +29,9 @@ def _get_embed_model():
if _embed_model is None:
from sentence_transformers import SentenceTransformer
logger.info("Chargement du modèle d'embedding pour la recherche...")
_embed_model = SentenceTransformer("dangvantuan/sentence-camembert-large", device="cpu")
import torch
_device = "cuda" if torch.cuda.is_available() else "cpu"
_embed_model = SentenceTransformer("dangvantuan/sentence-camembert-large", device=_device)
_embed_model.max_seq_length = 512
return _embed_model