[PYTHON] Apprentissage automatique avec docker (42) Programmation PyTorch pour l'apprentissage en profondeur par Ian Pointer

Programming PyTorch for Deep Learning By Ian Pointer

https://www.oreilly.com/library/view/programming-pytorch-for/9781492045342/

informel? https://github.com/falloutdurham/pytorchupandrunning

OS

$ docker run -p 8888:8888 -it kaizenjapan/pytorch /bin/bash

docker/anaconda

 # jupyter notebook --ip=0.0.0.0 --allow-root

Dans le navigateur localhost:8888 Ouvrir avec Entrez l'identifiant.

Procédure pour construire

OS

$ docker run -p 8888:8888 -it continuumio/anaconda3 /bin/bash

ubuntu/anaconda

# apt update; apt -y upgrade
# apt install vim git sudo apt-utils
# pip install --upgrade pip
# git clone https://github.com/falloutdurham/pytorchupandrunning.git
# conda install pytorch torchvision -c pytorch
Collecting package metadata (current_repodata.json): done
Solving environment: done

## Package Plan ##

  environment location: /opt/conda

  added / updated specs:
    - pytorch
    - torchvision


The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    conda-4.7.12               |           py37_0         3.0 MB
    cudatoolkit-10.1.243       |       h6bb024c_0       347.4 MB
    ninja-1.9.0                |   py37hfd86e86_0         1.2 MB
    pytorch-1.3.0              |py3.7_cuda10.1.243_cudnn7.6.3_0       463.7 MB  pytorch
    torchvision-0.4.1          |       py37_cu101        10.9 MB  pytorch
    ------------------------------------------------------------
                                           Total:       826.3 MB

The following NEW packages will be INSTALLED:

  cudatoolkit        pkgs/main/linux-64::cudatoolkit-10.1.243-h6bb024c_0
  ninja              pkgs/main/linux-64::ninja-1.9.0-py37hfd86e86_0
  pytorch            pytorch/linux-64::pytorch-1.3.0-py3.7_cuda10.1.243_cudnn7.6.3_0
  torchvision        pytorch/linux-64::torchvision-0.4.1-py37_cu101

The following packages will be UPDATED:

  conda                                       4.7.10-py37_0 --> 4.7.12-py37_0


Proceed ([y]/n)? y


Downloading and Extracting Packages
torchvision-0.4.1    | 10.9 MB   | ############################################################### | 100% 
pytorch-1.3.0        | 463.7 MB  | ############################################################### | 100% 
conda-4.7.12         | 3.0 MB    | ########################################################################### | 100% 
ninja-1.9.0          | 1.2 MB    | ########################################################################### | 100% 
cudatoolkit-10.1.243 | 347.4 MB  | ########################################################################### | 100% 
Preparing transaction: done
Verifying transaction: done
Executing transaction: done

Enregistrer dans le hub

Référence

GET STARTED pytorch https://pytorch.org/get-started/locally/

Je souhaite utiliser Pytorch dans un environnement virtuel https://qiita.com/nozomi254/items/b536883b1b46ae20aeae

Historique du document

ver. 0.01 Première version 20191105 ver. 0.02 Ajout d'images, etc. 20191106