.

app/main.py

@@ -92,7 +92,7 @@ def load_model(device: str):
     if device == "cpu":
         # 屏蔽 GPU，让后续 torch / BGEM3 都认不出 CUDA
         os.environ["CUDA_VISIBLE_DEVICES"] = ""
-        
+
     # Simple DataParallel for multi-GPU inference
     if device.startswith("cuda") and torch.cuda.device_count() > 1:
         logger.info(
@@ -179,16 +179,14 @@ logger.info("Using SAFE_MIN_FREE_MB = %d MB", SAFE_MIN_FREE_MB)
 async def warm_up_mp_pool():
     try:
         if DEVICE.startswith("cuda"):
-            logger.info("Warm-up (GPU) → 建多进程池")
-            _ = model.encode(["warmup"], return_dense=True)          # GPU
+            logger.info("Warm-up (GPU) → 预生成多进程池")
+            _ = model.encode(["warmup"], return_dense=True)
         else:
             logger.info("Warm-up (CPU) → 单进程初始化")
-            # 双保险：彻底把 BGEM3 的设备表改成只含 CPU
             if hasattr(model, "devices"):
-                model.devices = ["cpu"]
+                model.devices = ["cpu"]   # 彻底屏蔽 GPU
             model.device = "cpu"
-            _ = model.encode(["warmup"], return_dense=True,
-                             num_processes=1)                       # 不再 fork
+            _ = model.encode(["warmup"], return_dense=True)  # ← 删掉 num_processes
     except Exception as e:
         logger.warning("Warm-up failed: %s —— 首条请求时再退避", e)
 
@@ -209,12 +207,9 @@ def _encode(texts: List[str]):
     def _worker(t, q):
         try:
             if DEVICE.startswith("cuda"):
-                # 正常 GPU 跑，多进程池照旧
                 out = model.encode(t, return_dense=True)
             else:
-                # 已经 fallback 到 CPU：禁用进程池，禁止再向 GPU 拷权重
-                out = model.encode(t, return_dense=True,
-                                   num_processes=1)
+                out = model.encode(t, return_dense=True)  # ← 同样不传 num_processes
             q.put(("ok", out))
         except Exception as e:
             q.put(("err", str(e)))