更新

models/LatentSync/scripts/inference.py

@@ -14,6 +14,12 @@
 
 import argparse
 import os
+
+# --- 性能优化: 限制 CPU 线程数 ---
+os.environ["OMP_NUM_THREADS"] = "8" 
+os.environ["MKL_NUM_THREADS"] = "8"
+os.environ["TORCH_NUM_THREADS"] = "8"
+
 from omegaconf import OmegaConf
 import torch
 from diffusers import AutoencoderKL, DDIMScheduler

models/LatentSync/scripts/server.py

@@ -37,6 +37,14 @@ def load_gpu_config():
 
 load_gpu_config()
 
+# --- 性能优化: 限制 CPU 线程数 ---
+# 防止 PyTorch 默认占用所有 CPU 核心 (56线程) 导致系统卡顿
+# 预留资源给 Backend, Frontend 和 SSH
+os.environ["OMP_NUM_THREADS"] = "8" 
+os.environ["MKL_NUM_THREADS"] = "8"
+os.environ["TORCH_NUM_THREADS"] = "8"
+print("⚙️  已限制 PyTorch CPU 线程数为 8，防止系统卡顿")
+
 import torch
 from contextlib import asynccontextmanager
 from fastapi import FastAPI, HTTPException