Update the installation guide

Author: BillHuang2001

docs/source/locale/zh_CN/tutorial/tutorial_part2.md

@@ -76,9 +76,16 @@ Linux 系统（如 Ubuntu）下安装 EvoX 相对直接，大部分情况下可
 ```
 
 ```{tip}
-在 Linux 上，您只需安装驱动即可，通常不需要安装 CUDA 或其他依赖项。这些库已经包含在通过 pip 安装的 PyTorch 中。
+你很可能只需要安装驱动程序，而**不需要**单独安装 CUDA 或其他依赖项。
+这些库已经通过 pip 安装的 PyTorch 包含在内。
 ```
 
+```{tip}
+所需的驱动程序版本取决于你的硬件。如果你使用的是较新的 NVIDIA 显卡，通常推荐安装最新版驱动程序。
+为了获得更好的兼容性以及使用最新驱动程序，通常建议使用较新的 Linux 发行版（例如 Ubuntu 25.04 而不是 22.04）。
+```
+
+
 3. **安装 PyTorch**：同 Windows 类似，先安装 PyTorch 以确保硬件加速正常。可以参考 [PyTorch 官方指南](https://pytorch.org)
 
 4. **安装 EvoX**：运行
@@ -95,12 +102,23 @@ Linux 系统（如 Ubuntu）下安装 EvoX 相对直接，大部分情况下可
 
    这将同时安装可视化模块和 Brax 等神经进化相关依赖 ([EvoX Installation Guide](#EvoX Installation Guide))。您也可以根据需要选择 extras，比如只安装 `vis` 或 `neuroevolution`。
 
+  ````{note}
+  某些软件包可能需要额外的系统依赖项。如果是这种情况，安装程序会提示类似以下的消息：
+
+  ```console
+  error: Microsoft Visual C++ 14.0 or greater is required. Get it with "Microsoft C++ Build Tools": https://visualstudio.microsoft.com/visual-cpp-build-tools/
+  ```
+
+  遇到此类提示时，请按照提供的指引安装所需的系统依赖项后再继续操作。
+  ````
+
+
 #### 使用容器安装 (Docker, Podman)
 
 对于 AMD GPU 用户或希望隔离环境的用户，官方建议使用 Docker 镜像。例如，使用带 ROCm 的 PyTorch 官方Docker 镜像可以避免繁琐的环境配置。执行类似如下的命令运行容器：
 
 ```bash
-docker run -it --gpus all --shm-size=8g pytorch/pytorch:rocm5.4_ubuntu20.04
+docker run -it --gpus all --shm-size=8g rocm/pytorch:latest
 ```
 
 然后在容器内安装 EvoX（同上面的pip步骤）。这种方式可以方便地获取GPU加速支持。

docs/source/tutorial/tutorial_part2.md

@@ -59,6 +59,18 @@ To manually install EvoX:
    pip install "evox[neuroevolution]" # Neuroevolution support
    ```
 
+
+  ````{note}
+  Some packages may require additional system dependencies. If this is the case, the installer will prompt you with a message like the following:
+
+  ```console
+  error: Microsoft Visual C++ 14.0 or greater is required. Get it with "Microsoft C++ Build Tools": https://visualstudio.microsoft.com/visual-cpp-build-tools/
+  ```
+
+  When you encounter such messages, follow the provided instructions to install the necessary dependencies before proceeding.
+  ````
+
+
 ### Installation on Linux
 
 Installing EvoX on Linux (e.g., Ubuntu) is straightforward and mostly handled via `pip`.
@@ -72,7 +84,13 @@ On WSL, **do not** install NVIDIA drivers inside the Linux subsystem—install t
 ```
 
 ```{tip}
-It's very likely that you only need to install the driver, but do NOT need to install CUDA or other dependencies. Those libraries are already included in the PyTorch installation via pip.
+It's very likely that you only need to install the driver, but do NOT need to install CUDA or other dependencies.
+Those libraries are already included in the PyTorch installation via pip.
+```
+
+```{tip}
+The required driver version depends on your hardware. If you have a recent NVIDIA GPU, using the latest driver version is often the best choice.
+To ensure better compatibility and access to the latest drivers, it's generally a good idea to use a newer Linux distribution (e.g., Ubuntu 25.04 instead of 22.04).
 ```
 
 1. **Install PyTorch**: As on Windows, install based on hardware. Refer to the [PyTorch official guide](https://pytorch.org).
@@ -96,7 +114,7 @@ It's very likely that you only need to install the driver, but do NOT need to in
 For AMD GPU users or those seeking environment isolation, Docker is recommended. For example, using the official PyTorch Docker image with ROCm:
 
 ```bash
-docker run -it --gpus all --shm-size=8g pytorch/pytorch:rocm5.4_ubuntu20.04
+docker run -it --gpus all --shm-size=8g rocm/pytorch:latest
 ```
 
 Inside the container, install EvoX as usual using `pip`.