Upgrade to PyTorch 1.13 (#42)

Author: anibali

.github/workflows/publish_1.11.0-cuda11.5-ubuntu20.04.yml

@@ -19,5 +19,5 @@ jobs:
         username: ${{ secrets.DOCKER_USERNAME }}
         password: ${{ secrets.DOCKER_PASSWORD }}
         repository: anibali/pytorch
-        tags: 1.11.0-cuda11.5-ubuntu20.04,1.11.0-cuda11.5,latest
+        tags: 1.11.0-cuda11.5-ubuntu20.04,1.11.0-cuda11.5
         path: dockerfiles/1.11.0-cuda11.5-ubuntu20.04

.github/workflows/publish_1.10.0-cuda11.3-ubuntu20.04.yml .github/workflows/publish_1.13.0-cuda11.8-ubuntu22.04.yml

@@ -1,12 +1,12 @@
-name: Publish anibali/pytorch:1.10.0-cuda11.3-ubuntu20.04
+name: Publish anibali/pytorch:1.13.0-cuda11.8-ubuntu22.04
 
 on:
   push:
     branches:
     - master
     paths:
-    - dockerfiles/1.10.0-cuda11.3-ubuntu20.04/*
-    - .github/workflows/publish_1.10.0-cuda11.3-ubuntu20.04.yml
+    - dockerfiles/1.13.0-cuda11.8-ubuntu22.04/*
+    - .github/workflows/publish_1.13.0-cuda11.8-ubuntu22.04.yml
   workflow_dispatch:
 
 jobs:
@@ -19,5 +19,5 @@ jobs:
         username: ${{ secrets.DOCKER_USERNAME }}
         password: ${{ secrets.DOCKER_PASSWORD }}
         repository: anibali/pytorch
-        tags: 1.10.0-cuda11.3-ubuntu20.04,1.10.0-cuda11.3
-        path: dockerfiles/1.10.0-cuda11.3-ubuntu20.04
+        tags: 1.13.0-cuda11.8-ubuntu22.04,1.13.0-cuda11.8,latest
+        path: dockerfiles/1.13.0-cuda11.8-ubuntu22.04

.github/workflows/publish_1.10.0-nocuda-ubuntu20.04.yml .github/workflows/publish_1.13.0-nocuda-ubuntu22.04.yml

@@ -1,12 +1,12 @@
-name: Publish anibali/pytorch:1.10.0-nocuda-ubuntu20.04
+name: Publish anibali/pytorch:1.13.0-nocuda-ubuntu22.04
 
 on:
   push:
     branches:
     - master
     paths:
-    - dockerfiles/1.10.0-nocuda-ubuntu20.04/*
-    - .github/workflows/publish_1.10.0-nocuda-ubuntu20.04.yml
+    - dockerfiles/1.13.0-nocuda-ubuntu22.04/*
+    - .github/workflows/publish_1.13.0-nocuda-ubuntu22.04.yml
   workflow_dispatch:
 
 jobs:
@@ -19,5 +19,5 @@ jobs:
         username: ${{ secrets.DOCKER_USERNAME }}
         password: ${{ secrets.DOCKER_PASSWORD }}
         repository: anibali/pytorch
-        tags: 1.10.0-nocuda-ubuntu20.04,1.10.0-nocuda
-        path: dockerfiles/1.10.0-nocuda-ubuntu20.04
+        tags: 1.13.0-nocuda-ubuntu22.04,1.13.0-nocuda
+        path: dockerfiles/1.13.0-nocuda-ubuntu22.04

.github/workflows/publish_1.7.0-cuda11.0-ubuntu20.04.yml

@@ -0,0 +1,26 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
+
+## 2022-12-12
+
+### Added
+
+- The following new Docker images have been added:
+  - `1.13.0-cuda11.8-ubuntu22.04`
+  - `1.13.0-nocuda-ubuntu22.04`
+- New Docker images include [Micromamba](https://mamba.readthedocs.io/en/latest/user_guide/micromamba.html).
+
+### Changed
+
+- New Dockerfiles use an explicit [Conda lock file](https://github.com/conda-incubator/conda-lock) for better reproducability of image builds instead of installing from `environment.yml` directly. This generally won't affect users of images, but it does make it easier to see exactly which Conda packages are included in each image (see [dockerfiles/1.13.0-nocuda-ubuntu22.04/conda-linux-64.lock](dockerfiles/1.13.0-nocuda-ubuntu22.04/conda-linux-64.lock), for example). It is still possible to install new packages in derived images/containers using an `environment.yml` file.
+
+### Deprecated
+
+- The following Docker images will remain available on Docker Hub but are no longer supported:
+  - `1.7.0-cuda11.0-ubuntu20.04`
+  - `1.8.1-cuda11.1-ubuntu20.04`
+  - `1.10.0-nocuda-ubuntu20.04`
+  - `1.10.0-cuda11.3-ubuntu20.04`

.github/workflows/publish_1.8.1-cuda11.1-ubuntu20.04.yml

@@ -1,20 +1,19 @@
-## PyTorch Docker image
+# PyTorch Docker image
 
-[![Docker Automated build](https://img.shields.io/docker/automated/anibali/pytorch)](https://hub.docker.com/r/anibali/pytorch/)
-[![Docker Automated build](https://img.shields.io/docker/image-size/anibali/pytorch/latest)](https://hub.docker.com/r/anibali/pytorch/)
-[![Docker Automated build](https://img.shields.io/docker/pulls/anibali/pytorch)](https://hub.docker.com/r/anibali/pytorch/)
-[![Docker Automated build](https://img.shields.io/docker/v/anibali/pytorch/latest)](https://hub.docker.com/r/anibali/pytorch/)
+[![Docker image version](https://img.shields.io/docker/image-size/anibali/pytorch/latest)](https://hub.docker.com/r/anibali/pytorch/)
+[![Docker image pulls](https://img.shields.io/docker/pulls/anibali/pytorch)](https://hub.docker.com/r/anibali/pytorch/)
+[![Docker image size](https://img.shields.io/docker/v/anibali/pytorch/latest)](https://hub.docker.com/r/anibali/pytorch/)
 
 Ubuntu + PyTorch + CUDA (optional)
 
 
-### Requirements
+## Requirements
 
 In order to use this image you must have Docker Engine installed. Instructions
 for setting up Docker Engine are
 [available on the Docker website](https://docs.docker.com/engine/installation/).
 
-#### CUDA requirements
+### CUDA requirements
 
 If you have a CUDA-compatible NVIDIA graphics card, you can use a CUDA-enabled
 version of the PyTorch image to enable hardware acceleration. I have only
@@ -33,21 +32,21 @@ access within Docker containers. This can be found at
 [NVIDIA/nvidia-docker](https://github.com/NVIDIA/nvidia-docker).
 
 
-### Prebuilt images
+## Prebuilt images
 
 Prebuilt images are available on Docker Hub under the name
 [anibali/pytorch](https://hub.docker.com/r/anibali/pytorch/).
 
-For example, you can pull an image with PyTorch 1.8.1 and CUDA 11.1 using:
+For example, you can pull an image with PyTorch 1.13.0 and CUDA 11.8 using:
 
 ```bash
-$ docker pull anibali/pytorch:1.8.1-cuda11.1
+$ docker pull anibali/pytorch:1.13.0-cuda11.8
 ```
 
 
-### Usage
+## Usage
 
-#### Running PyTorch scripts
+### Running PyTorch scripts
 
 It is possible to run PyTorch programs inside a container using the
 `python3` command. For example, if you are within a directory containing
@@ -77,7 +76,7 @@ Here's a description of the Docker command-line options shown above:
 * `--volume="$PWD:/app"`: Mounts the current working directory into the container.
   The default working directory inside the container is `/app`. Optional.
 
-#### Running graphical applications
+### Running graphical applications
 
 If you are running on a Linux host, you can get code running inside the Docker
 container to display graphics using the host X server (this allows you to use
@@ -104,18 +103,18 @@ docker run --rm -it --init \
   anibali/pytorch python3 -c "import tkinter; tkinter.Tk().mainloop()"
 ```
 
-#### Deriving your own images
+### Deriving your own images
 
 The recommended way of adding additional dependencies to an image is to create
 your own Dockerfile using one of the PyTorch images from this project as a base.
 
 For example, let's say that you require OpenCV and wish to work with PyTorch
-1.8.1. You can create your own Dockerfile using
-`anibali/pytorch:1.8.1-cuda11.1-ubuntu20.04` as the base image and install
+1.13.0. You can create your own Dockerfile using
+`anibali/pytorch:1.13.0-cuda11.8-ubuntu22.04` as the base image and install
 OpenCV using additional build steps:
 
 ```dockerfile
-FROM anibali/pytorch:1.8.1-cuda11.1-ubuntu20.04
+FROM anibali/pytorch:1.13.0-cuda11.8-ubuntu22.04
 
 # Set up time zone.
 ENV TZ=UTC
@@ -131,7 +130,7 @@ RUN pip install opencv-python==4.5.1.48
 ```
 
 
-### Development and contributing
+## Development and contributing
 
 The Dockerfiles in the `dockerfiles/` directory are automatically generated by
 the `manager.py` script using details in `images.yml` and the templates in

CHANGELOG.md

@@ -0,0 +1,47 @@
+FROM nvidia/cuda:11.7.1-base-ubuntu22.04
+
+# Remove any third-party apt sources to avoid issues with expiring keys.
+RUN rm -f /etc/apt/sources.list.d/*.list
+
+# Install some basic utilities.
+RUN apt-get update && apt-get install -y \
+    curl \
+    ca-certificates \
+    sudo \
+    git \
+    bzip2 \
+    libx11-6 \
+ && rm -rf /var/lib/apt/lists/*
+
+# Create a working directory.
+RUN mkdir /app
+WORKDIR /app
+
+# Create a non-root user and switch to it.
+RUN adduser --disabled-password --gecos '' --shell /bin/bash user \
+ && chown -R user:user /app
+RUN echo "user ALL=(ALL) NOPASSWD:ALL" > /etc/sudoers.d/90-user
+USER user
+
+# All users can use /home/user as their home directory.
+ENV HOME=/home/user
+RUN mkdir $HOME/.cache $HOME/.config \
+ && chmod -R 777 $HOME
+
+# Download and install Micromamba.
+RUN curl -sL https://micro.mamba.pm/api/micromamba/linux-64/1.1.0 \
+  | sudo tar -xvj -C /usr/local bin/micromamba
+ENV MAMBA_EXE=/usr/local/bin/micromamba \
+    MAMBA_ROOT_PREFIX=/home/user/micromamba \
+    CONDA_PREFIX=/home/user/micromamba \
+    PATH=/home/user/micromamba/bin:$PATH
+
+# Set up the base Conda environment by installing PyTorch and friends.
+COPY conda-linux-64.lock /app/conda-linux-64.lock
+RUN micromamba create -qy -n base -f /app/conda-linux-64.lock \
+ && rm /app/conda-linux-64.lock \
+ && micromamba shell init --shell=bash --prefix="$MAMBA_ROOT_PREFIX" \
+ && micromamba clean -qya
+
+# Set the default command to python3.
+CMD ["python3"]