在 VPS 上搭建私有 AI 编程助手：Continue.dev + Ollama 完全指南

# SSH 到你的 VPS
ssh root@你的VPS-IP

# 一键安装 Ollama
curl -fsSL https://ollama.ai/install.sh | sh

# 验证安装
ollama --version

# DeepSeek Coder V2（推荐，代码能力最强）
ollama pull deepseek-coder-v2:16b

# Qwen2.5 Coder 7B（轻量级，响应快）
ollama pull qwen2.5-coder:7b

# Llama 3.1 8B（通用对话辅助）
ollama pull llama3.1:8b

# 查看已下载模型
ollama list

# 编辑 systemd 服务配置
systemctl edit ollama

# 添加以下内容（覆盖默认配置）：

[Service]
Environment="OLLAMA_HOST=0.0.0.0:11434"

# 重启 Ollama
systemctl daemon-reload
systemctl restart ollama

# 验证监听状态
ss -tlnp | grep 11434

# 安装 ufw
apt install -y ufw

# 只允许你的家庭/办公室 IP 访问
ufw allow from 你的本机IP到 any port 11434 proto tcp
ufw enable

# 或者使用 WireGuard VPN（更推荐）
# 通过 VPN 连接后，Ollama 只监听在 VPN 接口上

{
  "models": [
    {
      "title": "DeepSeek Coder V2 (VPS)",
      "provider": "ollama",
      "model": "deepseek-coder-v2:16b",
      "apiBase": "http://你的VPS-IP:11434",
      "completionOptions": {
        "temperature": 0.1,
        "topP": 0.9,
        "maxTokens": 4096
      }
    },
    {
      "title": "Qwen2.5 Coder 7B (VPS)",
      "provider": "ollama",
      "model": "qwen2.5-coder:7b",
      "apiBase": "http://你的VPS-IP:11434",
      "completionOptions": {
        "temperature": 0.1,
        "topP": 0.9,
        "maxTokens": 2048
      }
    }
  ],
  "tabAutocompleteModel": {
    "title": "Qwen2.5 Coder (Tab Auto)",
    "provider": "ollama",
    "model": "qwen2.5-coder:7b",
    "apiBase": "http://你的VPS-IP:11434"
  },
  "slashCommands": [
    {
      "name": "edit",
      "description": "编辑选中代码"
    },
    {
      "name": "comment",
      "description": "为代码添加注释"
    },
    {
      "name": "optimize",
      "description": "优化代码性能"
    },
    {
      "name": "explain",
      "description": "解释代码逻辑"
    }
  ]
}

# 安装 Nginx
apt install -y nginx

# 创建配置
cat > /etc/nginx/sites-available/ollama-proxy << 'EOF'
server {
    listen 443 ssl;
    server_name ollama.你的域名.com;

    ssl_certificate /etc/letsencrypt/live/ollama.你的域名.com/fullchain.pem;
    ssl_certificate_key /etc/letsencrypt/live/ollama.你的域名.com/privkey.pem;

    location / {
        # API 密钥验证
        if ($http_x_api_key != "你的密钥") {
            return 401;
        }

        proxy_pass http://127.0.0.1:11434;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
    }
}
EOF

# 启用配置
ln -s /etc/nginx/sites-available/ollama-proxy /etc/nginx/sites-enabled/
nginx -t && systemctl reload nginx

# 在 GPU VPS 上安装 NVIDIA 驱动
apt install -y nvidia-driver-545 nvidia-utils-545

# 验证 GPU
nvidia-smi

# Ollama 自动检测 GPU，无需额外配置
# 运行 16B 模型时 GPU 推理比 CPU 快 5-10 倍

upstream ollama_backend {
    server vps1:11434 weight=3;
    server vps2:11434 weight=1;
}

# 创建 16GB 内存盘（需要足够 RAM）
mount -t tmpfs -o size=16G tmpfs /mnt/ramdisk

# 将模型链接到内存盘
ln -s /mnt/ramdisk/blobs /usr/share/ollama/.ollama/models/blobs

def fibonacci(n: int) -> list[int]:
    """Generate first n Fibonacci numbers."""
    if n <= 0:
        return []
    fib = [0, 1]
    for _ in range(2, n):
        fib.append(fib[-1] + fib[-2])
    return fib[:n]