Skip to content

Commit 7ea2ed4

Browse files
tianshuo78520atianshuo78520a
authored
Update en install index (#7285)
* Update en install index * Update * Update * Update
1 parent 3bb5c27 commit 7ea2ed4

File tree

1 file changed

+48
-131
lines changed

1 file changed

+48
-131
lines changed

‎docs/install/index_en.rst

Lines changed: 48 additions & 131 deletions
Original file line numberDiff line numberDiff line change
@@ -9,8 +9,7 @@
99
Important updates
1010
----------------------
1111

12-
* Add support for python3.12, and no longer supports python3.7
13-
* Add support for CUDA 12.0, and no longer supports CUDA 10.2
12+
* Paddle supports user installation without depending on CUDA and cuDNN, and automatically handles the installation of CUDA and cuDNN.
1413

1514

1615
------------------------
@@ -23,10 +22,10 @@ The manuals will guide you to build and install PaddlePaddle on your 64-bit desk
2322
1. Operating system requirements:
2423
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
2524

26-
* Windows 7 / 8 / 10, Pro/Enterprise
27-
* Ubuntu 18.04 / 20.04
28-
* CentOS 7
29-
* macOS 10.x/11.x/12.x/13.x/14.x
25+
* Windows 7 / 8 / 10 / 11, Pro/Enterprise
26+
* Ubuntu 20.04 / 22.04
27+
* almalinux 8
28+
* macOS 12.x/13.x/14.x/15.x
3029
* 64-bit operating system is required
3130

3231
2. Processor requirements:
@@ -38,180 +37,99 @@ The manuals will guide you to build and install PaddlePaddle on your 64-bit desk
3837
3. Version requirements of python and pip:
3938
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
4039

41-
* Python requires version 3.8/3.9/3.10/3.11/3.12
40+
* Python requires version 3.8/3.9/3.10/3.11/3.12/3.13
4241
* Python needs pip, and pip requires version 20.2.2 or above
4342
* Python and pip requires 64-bit
4443

45-
4. PaddlePaddle's support for GPU:
46-
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
44+
**First Installation Method: Using pip for installation**
45+
You can choose any of the four methods: "Using pip for installation", "Using conda for installation", "Using docker for installation", "Compiling from source code for installation".
4746

48-
* Currently, **PaddlePaddle** supports **CUDA** driver of **NVIDIA** graphics card and **ROCm** driver of **AMD** card.
49-
* You need to install `cuDNN <https://docs.nvidia.com/deeplearning/sdk/cudnn-install/>`_ , and version 7.6+ is required(For CUDA11)
50-
* If you need GPU multi-card mode, you need to install `NCCL 2 <https://developer.nvidia.com/nccl/>`_
47+
This section will introduce the installation method using pip.
5148

52-
* Only Ubuntu/CentOS support NCCL 2
53-
* You need to install `CUDA <https://docs.nvidia.com/cuda/cuda-installation-guide-windows/>`_ , depending on your system, there are different requirements for CUDA version:
49+
1.You need to ensure that your operating system meets the requirements listed above.
5450

55-
* Windows install GPU version
51+
2.You need to ensure that your processor meets the requirements listed above.
5652

57-
* Windows 7 / 8 / 10 support CUDA 11.0/11.2/11.6/11.8/12.0 single-card mode
58-
* don't support install using **nvidia-docker**
59-
* Ubuntu install GPU version
53+
3.Ensure that the Python where you need to install PaddlePaddle is in your expected location, as your computer may have multiple Pythons.
6054

61-
* Ubuntu 18.04 supports CUDA (11.0 - 12.0)
62-
* Ubuntu 20.04 supports CUDA (11.0 - 12.0)
63-
* If you install using **nvidia-docker** , it supports CUDA 11.2/11.7/12.0
64-
* CentOS install GPU version
55+
Use the following command to output the Python path, depending on your environment you may need to replace all the python in the command line in the instructions with the specific Python path.
6556

66-
* If you install using native **pip** :
57+
In the Windows environment, the command to output the Python path is:
6758

68-
* CentOS 7 supports CUDA (11.0 - 12.0)
69-
* If you compile and install using native source code:
59+
::
7060

71-
* CentOS 7 supports CUDA (11.0 - 12.0)
72-
* If you install using **nvidia-docker** , CentOS 7 supports CUDA 11.2/11.7/12.0
73-
* macOS isn't supported: PaddlePaddle has no GPU support in Mac OS platform
61+
where python
7462

75-
Please make sure your environment meets the above conditions. If you have other requirements, please refer to `Appendix <Tables_en.html#ciwhls-release>`_ .
63+
In the macOS/Linux environment, the command to output the Python path is:
7664

77-
5. PaddlePaddle's support for NCCL:
78-
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
65+
::
7966

80-
* Support for Windows
67+
which python
8168

82-
* not support NCCL
83-
* Support for Ubuntu
69+
4.Check the Python version
8470

85-
* Ubuntu 18.04:
71+
Use the following command to confirm it is 3.8/3.9/3.10/3.11/3.12/3.13
8672

87-
* support NCCL v2.4.7 / v2.16.5 under CUDA11
88-
* Ubuntu 20.04:
89-
90-
* support v2.4.7 / 2.16.5 under CUDA11
91-
* Support for CentOS
92-
93-
* CentOS 7:
94-
95-
* support NCCL v2.4.7-v2.16.5 under CUDA11
96-
* Support for macOS
97-
98-
* not support NCCL
99-
100-
101-
The first way to install: use pip to install
102-
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
103-
104-
You can choose any of the four ways to install: "use pip to install", "use Conda to install", "use Docker to install", "compiling from the source code"
105-
106-
This section describes how to use pip to install.
107-
108-
1. You need to confirm that your operating system meets the requirements listed above
109-
110-
2. You need to confirm that your processor meets the requirements listed above
111-
112-
3. Confirm that the Python where you need to install PaddlePaddle is your expected location, because your computer may have multiple Python
113-
114-
Use the following command to output Python path. Depending on your environment, you may need to replace Python in all command lines in the description with specific Python path
115-
116-
In the Windows environment, the command to output Python path is:
117-
118-
::
119-
120-
where python
121-
122-
In the macOS/Linux environment, the command to output Python path is:
123-
124-
::
125-
126-
which python
127-
128-
129-
4. Check the version of Python
130-
131-
Confirm the Python is 3.8/3.9/3.10/3.11/3.12 using command
13273
::
13374

13475
python --version
13576

136-
5. Check the version of pip and confirm it is 20.2.2 or above
77+
5.Check the pip version, confirm it is 20.2.2+
13778

13879
::
13980

14081
python -m ensurepip
14182
python -m pip --version
14283

143-
144-
6. Confirm that Python and pip is 64 bit，and the processor architecture is x86_64(or called x64、Intel 64、AMD64)architecture. Currently. The first line below outputs "64bit", and the second line outputs "x86_64", "x64" or "AMD64" :
84+
6.Confirm that Python and pip are 64 bit, and the processor architecture is x86_64 (also known as x64, Intel 64, AMD64). The first line of the following output is "64bit", and the second line output is "x86_64", "x64" or "AMD64":
14585

14686
::
14787

14888
python -c "import platform;print(platform.architecture()[0]);print(platform.machine())"
14989

90+
7.If you want to use pip <https://pypi.org/project/pip/>_ to install PaddlePaddle, you can directly use the following command:
15091

151-
7. If you want to use `pip <https://pypi.org/project/pip/>`_ to install PaddlePaddle, you can use the command below directly:
152-
153-
(1). **CPU version** : If you only want to install CPU version, please refer to command below
154-
155-
Command to install CPU version is:
156-
::
157-
158-
python -m pip install paddlepaddle==2.6.0 -i https://mirror.baidu.com/pypi/simple
159-
160-
or
161-
162-
python -m pip install paddlepaddle==2.6.0 -i https://pypi.tuna.tsinghua.edu.cn/simple
163-
164-
165-
(2). **GPU version** : If you only want to install GPU version, please refer to command below
166-
167-
168-
Note:
169-
170-
* You need to confirm that your GPU meets the requirements listed above
171-
172-
::
173-
174-
python -m pip install paddlepaddle-gpu==2.6.0 -i https://mirror.baidu.com/pypi/simple
175-
176-
or
177-
178-
python -m pip install paddlepaddle-gpu==2.6.0 -i https://pypi.tuna.tsinghua.edu.cn/simple
92+
Note:
17993

94+
* If you want to install paddlepaddle, the version requires libstdc++.so.6 version greater than 3.4.25. To meet this requirement, you can choose to install GCC 8 or a higher GCC version, or upgrade the libstdc++ library separately. *
18095

181-
Please confirm that the Python where you need to install PaddlePaddle is your expected location, because your computer may have multiple Python. Depending on the environment, you may need to replace Python in all command lines in the instructions with Python 3 or specific Python path.
96+
(1). **CPU version** : If you just want to install the CPU version, please refer to the following command for installation
18297

183-
8. Verify installation
98+
The command to install the CPU version is:
99+
::
100+
python -m pip install --pre paddlepaddle -i https://www.paddlepaddle.org.cn/packages/nightly/cpu/
184101

185-
After the installation is complete, you can use `python` to enter the Python interpreter and then use `import paddle` and then `paddle.utils.run_check()` to verify that the installation was successful.
102+
(2). **GPU version** : If you want to use the GPU version, please refer to the following command for installation
186103

187-
If `PaddlePaddle is installed successfully!` appears, it means the installation was successful.
104+
The command to install the GPU cuda12.6 version is:
105+
::
106+
python -m pip install --pre paddlepaddle-gpu -i https://www.paddlepaddle.org.cn/packages/nightly/cu126/
188107

108+
The command to install the GPU cuda11.8 version is:
109+
::
110+
python -m pip install --pre paddlepaddle-gpu -i https://www.paddlepaddle.org.cn/packages/nightly/cu118/
189111

190-
9. For more information to help, please refer to:
112+
Please make sure that the Python where you need to install PaddlePaddle is in the expected location, as your computer may have multiple Pythons. Depending on your environment, you may need to replace all python in the command line in the instructions with the specific Python path.
191113

192-
`install under Ubuntu <pip/linux-pip_en.html>`_
114+
8.Verify the installation
193115

194-
`install under macOS <pip/macos-pip_en.html>`_
116+
Use python to enter the python interpreter, enter import paddle, then enter paddle.utils.run_check().
195117

196-
`install under Windows <pip/windows-pip_en.html>`_
118+
If PaddlePaddle is installed successfully! appears, it means you have successfully installed.
197119

120+
9.For more help information, please refer to:
198121

122+
`PIP Installation under Linux <pip/linux-pip.html>`_
199123

200-
The second way to install: compile and install with container
201-
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
124+
`PIP Installation under macOS <pip/macos-pip.html>`_
202125

203-
- We recommend that you use `NVIDIA PaddlePaddle Container <https://www.paddlepaddle.org.cn/documentation/docs/zh/install/install_NGC_PaddlePaddle_ch.html>`_ for your development environment installation.
204-
- Pros
205-
1. Lastest version of CUDA
206-
2. Newer verison of Ubuntu OS(18.04)
207-
3. Performance and development efficiency have been optimized by NVIDIA
126+
`PIP Installation under Windows <pip/windows-pip.html>`_
208127

128+
**Second Installation Method: Using Source Code Compilation for Installation**
209129

210-
The third way to install: compile and install with source code
211-
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
130+
- If you are just using PaddlePaddle, it is recommended to use **pip** for installation.
212131

213-
- If you use PaddlePaddle only, we suggest you installation methods **pip** to install.
214-
- If you need to develop PaddlePaddle, please refer to `compile from source code <compile/fromsource_en.html>`_
132+
- If you have a need to develop PaddlePaddle, please refer to: `Compiling from Source <compile/fromsource.html>`_
215133

216134
.. toctree::
217135
:hidden:
@@ -220,6 +138,5 @@ The third way to install: compile and install with source code
220138
conda/fromconda_en.rst
221139
docker/fromdocker_en.rst
222140
compile/fromsource_en.rst
223-
install_xpu_en.md
224141
install_NGC_PaddlePaddle_en.rst
225142
Tables_en.md

0 commit comments

Comments
 (0)