How to Install PyTorch on Ubuntu

admin

https://www.cherryservers.com/blog/how-to-install-pytorch-ubuntu

直接跳到

Conda: install PyTorch
这一部分

admin

执行
bash anaconda.sh
后出现以下信息

Do you accept the license terms? [yes|no]
[no] >>> yes

Anaconda3 will now be installed into this location:
/root/anaconda3

  - Press ENTER to confirm the location
  - Press CTRL-C to abort the installation
  - Or specify a different location below

[/root/anaconda3] >>> /mnt/anaconda3
PREFIX=/mnt/anaconda3
Unpacking payload ...

Installing base environment...


Downloading and Extracting Packages


Downloading and Extracting Packages

Preparing transaction: done
Executing transaction: -

    Installed package of scikit-learn can be accelerated using scikit-learn-intelex.
    More details are available here: https://intel.github.io/scikit-learn-intelex

    For example:

        $ conda install scikit-learn-intelex
        $ python -m sklearnex my_application.py


                                                                                                                                                                                                            done
installation finished.
Do you wish the installer to initialize Anaconda3
by running conda init? [yes|no]
[no] >>> yes
no change     /mnt/anaconda3/condabin/conda
no change     /mnt/anaconda3/bin/conda
no change     /mnt/anaconda3/bin/conda-env
no change     /mnt/anaconda3/bin/activate
no change     /mnt/anaconda3/bin/deactivate
no change     /mnt/anaconda3/etc/profile.d/conda.sh
no change     /mnt/anaconda3/etc/fish/conf.d/conda.fish
no change     /mnt/anaconda3/shell/condabin/Conda.psm1
no change     /mnt/anaconda3/shell/condabin/conda-hook.ps1
no change     /mnt/anaconda3/lib/python3.11/site-packages/xontrib/conda.xsh
no change     /mnt/anaconda3/etc/profile.d/conda.csh
modified      /root/.bashrc

==> For changes to take effect, close and re-open your current shell. <==

If you'd prefer that conda's base environment not be activated on startup,
   set the auto_activate_base parameter to false:

conda config --set auto_activate_base false

Thank you for installing Anaconda3!

admin

使用
nvidia-smi
查看当前CUDA Version
CUDA Version: 12.8

admin

根据这一篇
https://greenwebpage.com/community/how-to-install-pytorch-on-ubuntu-24-04/
Replace pytorch-cuda=11.7 with the correct version for your CUDA installation.
那我安装命令应该是
conda install pytorch torchvision torchaudio pytorch-cuda=12.8 -c pytorch -c nvidia

admin

出现如下错误提示

(base) root@study:/mnt/anaconda# conda install pytorch torchvision torchaudio pytorch-cuda=12.8 -c pytorch -c nvidia
Collecting package metadata (current_repodata.json): | DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): conda.anaconda.org:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): conda.anaconda.org:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): conda.anaconda.org:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): conda.anaconda.org:443                                                                                                                      / DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/main/noarch/current_repodata.json HTTP/1.1" 200 None                                                                                  / DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/r/linux-64/current_repodata.json HTTP/1.1" 200 None
DEBUG:urllib3.connectionpool:https://conda.anaconda.org:443 "GET /nvidia/noarch/current_repodata.json HTTP/1.1" 200 None                                                                                    - DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/main/linux-64/current_repodata.json HTTP/1.1" 200 None                                                                                | DEBUG:urllib3.connectionpool:https://conda.anaconda.org:443 "GET /nvidia/linux-64/current_repodata.json HTTP/1.1" 200 None
DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/r/noarch/current_repodata.json HTTP/1.1" 200 None                                                                                     - DEBUG:urllib3.connectionpool:https://conda.anaconda.org:443 "GET /pytorch/linux-64/current_repodata.json HTTP/1.1" 200 None                                                                                 / DEBUG:urllib3.connectionpool:https://conda.anaconda.org:443 "GET /pytorch/noarch/current_repodata.json HTTP/1.1" 200 None                                                                                   done
Solving environment: unsuccessful initial attempt using frozen solve. Retrying with flexible solve.
Collecting package metadata (repodata.json): \ DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): conda.anaconda.org:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): conda.anaconda.org:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): conda.anaconda.org:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): conda.anaconda.org:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): repo.anaconda.com:443                                                                                                                       | DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/main/noarch/repodata.json HTTP/1.1" 200 None                                                                                          / DEBUG:urllib3.connectionpool:https://conda.anaconda.org:443 "GET /nvidia/linux-64/repodata.json HTTP/1.1" 200 None                                                                                          \ DEBUG:urllib3.connectionpool:https://conda.anaconda.org:443 "GET /nvidia/noarch/repodata.json HTTP/1.1" 200 None
DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/r/noarch/repodata.json HTTP/1.1" 200 None                                                                                             - DEBUG:urllib3.connectionpool:https://conda.anaconda.org:443 "GET /pytorch/linux-64/repodata.json HTTP/1.1" 200 None                                                                                         \ DEBUG:urllib3.connectionpool:https://conda.anaconda.org:443 "GET /pytorch/noarch/repodata.json HTTP/1.1" 200 None
DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/r/linux-64/repodata.json HTTP/1.1" 200 None
DEBUG:urllib3.connectionpool:https://repo.anaconda.com:443 "GET /pkgs/main/linux-64/repodata.json HTTP/1.1" 200 None                                                                                        done
Solving environment: unsuccessful initial attempt using frozen solve. Retrying with flexible solve.

PackagesNotFoundError: The following packages are not available from current channels:

  - pytorch-cuda=12.8

Current channels:

  - https://conda.anaconda.org/pytorch/linux-64
  - https://conda.anaconda.org/pytorch/noarch
  - https://conda.anaconda.org/nvidia/linux-64
  - https://conda.anaconda.org/nvidia/noarch
  - https://repo.anaconda.com/pkgs/main/linux-64
  - https://repo.anaconda.com/pkgs/main/noarch
  - https://repo.anaconda.com/pkgs/r/linux-64
  - https://repo.anaconda.com/pkgs/r/noarch

To search for alternate channels that may provide the conda package you're
looking for, navigate to

    https://anaconda.org

and use the search bar at the top of the page.

admin

(base) root@study:/mnt/anaconda# conda env list

# conda environments:
#
base                  *  /mnt/anaconda3

(base) root@study:/mnt/anaconda# conda info

     active environment : base
    active env location : /mnt/anaconda3
            shell level : 1
       user config file : /root/.condarc
 populated config files :
          conda version : 23.7.2
    conda-build version : 3.26.0
         python version : 3.11.4.final.0
       virtual packages : __archspec=1=x86_64
                          __cuda=12.8=0
                          __glibc=2.39=0
                          __linux=6.8.0=0
                          __unix=0=0
       base environment : /mnt/anaconda3  (writable)
      conda av data dir : /mnt/anaconda3/etc/conda
  conda av metadata url : None
           channel URLs : https://repo.anaconda.com/pkgs/main/linux-64
                          https://repo.anaconda.com/pkgs/main/noarch
                          https://repo.anaconda.com/pkgs/r/linux-64
                          https://repo.anaconda.com/pkgs/r/noarch
          package cache : /mnt/anaconda3/pkgs
                          /root/.conda/pkgs
       envs directories : /mnt/anaconda3/envs
                          /root/.conda/envs
               platform : linux-64
             user-agent : conda/23.7.2 requests/2.31.0 CPython/3.11.4 Linux/6.8.0-55-generic ubuntu/24.04.2 glibc/2.39
                UID:GID : 0:0
             netrc file : None
           offline mode : False

admin

到这一页来看
https://pytorch.org/get-started/locally/
CUDA 12.8只能选择nightly pip安装，命令如下
pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128

2025/03/10
今天搜到这里
https://anaconda.org/nvidia/repo/installers?label=cuda-12.8.0
运行这个命令进行conda安装
conda install --channel "nvidia/label/cuda-12.8.0" package
使用上述命令不成功

admin

四个小时的安装时间，看起来是安装成功了，信息如下

Installing collected packages: pytorch-triton, nvidia-cusparselt-cu12, typing-extensions, sympy, nvidia-nvtx-cu12, nvidia-nvjitlink-cu12, nvidia-nccl-cu12, nvidia-curand-cu12, nvid ia-cufile-cu12, nvidia-cuda-runtime-cu12, nvidia-cuda-nvrtc-cu12, nvidia-cuda-cupti-cu12, nvidia-cublas-cu12, nvidia-cusparse-cu12, nvidia-cufft-cu12, nvidia-cudnn-cu12, nvidia-cus olver-cu12, torch, torchvision, torchaudio Attempting uninstall: typing-extensions Found existing installation: typing_extensions 4.7.1 Uninstalling typing_extensions-4.7.1: Successfully uninstalled typing_extensions-4.7.1 Attempting uninstall: sympy Found existing installation: sympy 1.11.1 Uninstalling sympy-1.11.1: Successfully uninstalled sympy-1.11.1 Attempting uninstall: torch Found existing installation: torch 2.0.1 Uninstalling torch-2.0.1: Successfully uninstalled torch-2.0.1 Successfully installed nvidia-cublas-cu12-12.8.3.14 nvidia-cuda-cupti-cu12-12.8.57 nvidia-cuda-nvrtc-cu12-12.8.61 nvidia-cuda-runtime-cu12-12.8.57 nvidia-cudnn-cu12-9.7.1.26 nvidia-cufft-cu12-11.3.3.41 nvidia-cufile-cu12-1.13.0.11 nvidia-curand-cu12-10.3.9.55 nvidia-cusolver-cu12-11.7.2.55 nvidia-cusparse-cu12-12.5.7.53 nvidia-cusparselt-cu12-0.6.3 nvidia-nccl-cu12-2.25.1 nvidia-nvjitlink-cu12-12.8.61 nvidia-nvtx-cu12-12.8.55 pytorch-triton-3.2.0+git4b3bb1f8 sympy-1.13.3 torch-2.7.0.dev20250308+cu128 torchaudio-2.6.0.dev20250308+cu128 torchvision-0.22.0.dev20250308+cu128 typing-extensions-4.12.2 WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv

admin

验证完毕

`(base) root@study:/mnt/anaconda# python
Python 3.11.4 (main, Jul 5 2023, 14:15:25) [GCC 11.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.

import torch
print(torch.version)
2.7.0.dev20250308+cu128
`

做一个备份，然后安装ComifyUI

admin

Error while loading conda entry point: anaconda-cloud-auth (cannot import name 'TypeIs' from 'typing_extensions' (/mnt/anaconda3/lib/python3.11/site-packages/typing_extensions.py))

admin

20250311
需要重装，目前看来Anacondabing不支持CUDA12.8
那么这次采用Pip: install PyTorch
现在有一个问题就是安装python的版本，安装的时候再看
参考这一篇
https://www.rosehosting.com/blog/how-to-install-python-on-ubuntu-24-04/
使用Install Python from Source方式