Update app.py

app.py

@@ -5,17 +5,17 @@ import torch
 import torch.nn.functional as F
 import os
 
-# 📁 Définir le dossier de cache autorisé
+# 📁 Définir le cache autorisé
 CACHE_DIR = "/data"
 os.environ["HF_HOME"] = CACHE_DIR
 os.environ["TRANSFORMERS_CACHE"] = CACHE_DIR
 
 app = FastAPI()
 
-# ✅ Charger le modèle nomic-embed-text-v1
+# ✅ Modèle avec custom code : activer trust_remote_code
 MODEL_NAME = "nomic-ai/nomic-embed-text-v1"
-tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, cache_dir=CACHE_DIR)
-model = AutoModel.from_pretrained(MODEL_NAME, cache_dir=CACHE_DIR)
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, cache_dir=CACHE_DIR, trust_remote_code=True)
+model = AutoModel.from_pretrained(MODEL_NAME, cache_dir=CACHE_DIR, trust_remote_code=True)
 
 class EmbedInput(BaseModel):
     text: str
@@ -25,6 +25,6 @@ async def embed_text(payload: EmbedInput):
     inputs = tokenizer(payload.text, return_tensors="pt", padding=True, truncation=True)
     with torch.no_grad():
         outputs = model(**inputs)
-        embeddings = outputs.last_hidden_state[:, 0]  # CLS token
+        embeddings = outputs.last_hidden_state[:, 0]
         normalized = F.normalize(embeddings, p=2, dim=1)
     return {"embedding": normalized[0].tolist()}