diff --git a/docs/get_started/installation/kunlunxin_xpu.md b/docs/get_started/installation/kunlunxin_xpu.md index 0a0cb3bdb3..7c50697356 100644 --- a/docs/get_started/installation/kunlunxin_xpu.md +++ b/docs/get_started/installation/kunlunxin_xpu.md @@ -110,44 +110,6 @@ bash build.sh The compiled outputs will be located in the ```FastDeploy/dist``` directory. -## 4. Python-only Quick Install (For Development) - -If you have already completed a full build (`bash build.sh 1`) and only modified Python files, you can use the python-only mode to quickly sync changes to `site-packages` **without recompiling C++ Custom Ops or rebuilding the wheel**. - -### Prerequisites - -A full build must have been completed at least once: - -```bash -bash build.sh 1 -``` - -### Usage - -```bash -# Argument 1: Must be 2 to enable python-only mode -# Argument 2 (optional): Python interpreter path (default: python) - -# Use default python -bash build.sh 2 - -# Use a specific python interpreter -bash build.sh 2 python3 - -# Use an absolute path to the interpreter -bash build.sh 2 /path/to/your/python -``` - -This command syncs `.py` files from the source tree directly into the installed `site-packages/fastdeploy/` directory via `rsync`. Compiled artifacts (`.so` files) are preserved. - -### When to Use - -| Scenario | Recommended Command | -|---|---| -| Only modified Python files | `bash build.sh 2` | -| Modified C++/CUDA code | `bash build.sh 1` | -| First-time build | `bash build.sh 1` | - ## Installation verification ```bash diff --git a/docs/get_started/installation/nvidia_gpu.md b/docs/get_started/installation/nvidia_gpu.md index 756d422c24..c59467175d 100644 --- a/docs/get_started/installation/nvidia_gpu.md +++ b/docs/get_started/installation/nvidia_gpu.md @@ -125,44 +125,6 @@ After the build completes, the operator binaries can be found in `FastDeploy/fas > - For custom architectures or modified operator logic, please use **source compilation (Section 4)**. > - You can check whether the precompiled wheel for a specific commit has been successfully built on the [FastDeploy CI Build Status Page](https://github.com/PaddlePaddle/FastDeploy/actions/workflows/ci_image_update.yml). -## 6. Python-only Quick Install (For Development) - -If you have already completed a full build (`bash build.sh 1 ...`) and only modified Python files, you can use the python-only mode to quickly sync changes to `site-packages` **without recompiling C++ Custom Ops or rebuilding the wheel**. - -### Prerequisites - -A full build must have been completed at least once: - -```shell -bash build.sh 1 python false [80,90] -``` - -### Usage - -```shell -# Argument 1: Must be 2 to enable python-only mode -# Argument 2 (optional): Python interpreter path (default: python) - -# Use default python -bash build.sh 2 - -# Use a specific python interpreter -bash build.sh 2 python3 - -# Use an absolute path to the interpreter -bash build.sh 2 /path/to/your/python -``` - -This command syncs `.py` files from the source tree directly into the installed `site-packages/fastdeploy/` directory via `rsync`. Compiled artifacts (`.so` files) are preserved. - -### When to Use - -| Scenario | Recommended Command | -|---|---| -| Only modified Python files | `bash build.sh 2` | -| Modified C++/CUDA code | `bash build.sh 1 python false [80,90]` | -| First-time build | `bash build.sh 1 python false [80,90]` | - ## Environment Verification After installation, verify the environment with this Python code: diff --git a/docs/zh/get_started/installation/kunlunxin_xpu.md b/docs/zh/get_started/installation/kunlunxin_xpu.md index d784216d23..b0e7f2a64f 100644 --- a/docs/zh/get_started/installation/kunlunxin_xpu.md +++ b/docs/zh/get_started/installation/kunlunxin_xpu.md @@ -111,40 +111,6 @@ bash build.sh 编译后的产物在 ```FastDeploy/dist``` 目录下。 -## 4. Python-only 快速安装 - -如果你已经完成过一次完整编译安装，后续仅修改了 Python 代码，可以使用 Python-only 模式将 Python 文件快速同步到 `site-packages`，**跳过 C++ 算子编译和 Wheel 打包**。 - -### 前提条件 - -必须先完成过一次完整编译安装（即执行过 `bash build.sh 1 ...`），确保 `site-packages` 中已有编译产物（`.so` 文件）。 - -### 使用方式 - -```bash -# 使用默认 python -bash build.sh 2 - -# 指定 Python 解释器 -bash build.sh 2 python3 - -# 使用完整路径指定 Python 解释器 -bash build.sh 2 /path/to/your/python -``` - -- 第 1 个参数 `2`：表示使用 Python-only 模式 -- 第 2 个参数（可选）：Python 解释器路径或命令名，默认为 `python` - -### 适用场景 - -| 场景 | 推荐方式 | -|------|----------| -| 仅修改 Python 文件 | `bash build.sh 2` | -| 修改了 C++/CUDA 代码 | `bash build.sh 1 ...`（完整编译） | -| 首次编译 | `bash build.sh 1 ...`（完整编译） | - -> **注意：** 该模式需要系统安装 `rsync`，若未安装请先执行 `apt-get install rsync` 或 `yum install rsync`。 - ## 验证是否安装成功 ```python diff --git a/docs/zh/get_started/installation/nvidia_gpu.md b/docs/zh/get_started/installation/nvidia_gpu.md index 15ef9d2779..8e989db536 100644 --- a/docs/zh/get_started/installation/nvidia_gpu.md +++ b/docs/zh/get_started/installation/nvidia_gpu.md @@ -131,40 +131,6 @@ bash build.sh 1 python false [90] 1 8a9e7b53af4a98583cab65e4b44e3265a93e56d2 > - 若希望自定义架构或修改算子逻辑，请使用 **源码编译方式（第4节）**。 > - 您可以在 FastDeploy CI 构建状态页面查看对应 commit 的预编译 whl 是否已构建成功。 -## 6. Python-only 快速安装 - -如果你已经完成过一次完整编译安装，后续仅修改了 Python 代码，可以使用 Python-only 模式将 Python 文件快速同步到 `site-packages`，**跳过 C++ 算子编译和 Wheel 打包**。 - -### 前提条件 - -必须先完成过一次完整编译安装（即执行过 `bash build.sh 1 ...`），确保 `site-packages` 中已有编译产物（`.so` 文件）。 - -### 使用方式 - -```bash -# 使用默认 python -bash build.sh 2 - -# 指定 Python 解释器 -bash build.sh 2 python3 - -# 使用完整路径指定 Python 解释器 -bash build.sh 2 /path/to/your/python -``` - -- 第 1 个参数 `2`：表示使用 Python-only 模式 -- 第 2 个参数（可选）：Python 解释器路径或命令名，默认为 `python` - -### 适用场景 - -| 场景 | 推荐方式 | -|------|----------| -| 仅修改 Python 文件 | `bash build.sh 2` | -| 修改了 C++/CUDA 代码 | `bash build.sh 1 python false "[80,90]"` | -| 首次编译 | `bash build.sh 1 python false "[80,90]"` | - -> **注意：** 该模式需要系统安装 `rsync`，若未安装请先执行 `apt-get install rsync` 或 `yum install rsync`。 - ## 环境检查 在安装 FastDeploy 后，通过如下 Python 代码检查环境的可用性