Update src/qa.py

src/qa.py

@@ -29,11 +29,14 @@ os.environ.update({
 })
 
 # ==========================================================
-# 2️⃣ Embedding Model (E5-large-v2)
+# 2️⃣ Embedding Model (E5-small-v2)
 # ==========================================================
 try:
-    _query_model = SentenceTransformer("intfloat/e5-large-v2", cache_folder=CACHE_DIR)
-    print("✅ Loaded embedding model: intfloat/e5-large-v2")
+    _query_model = SentenceTransformer(
+        "intfloat/e5-small-v2",   # ⚡ Faster, 384-dim embeddings
+        cache_folder=CACHE_DIR
+    )
+    print("✅ Loaded embedding model: intfloat/e5-small-v2 (fast mode)")
 except Exception as e:
     print(f"⚠️ Embedding load failed ({e}), using MiniLM fallback")
     _query_model = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2", cache_folder=CACHE_DIR)