• Logout
  • Profile
  • Login
  • Register
Machine learning
aise / PyTorch 0.4 Python 2.7 CUDA 9.1 Production

A pre-configured and fully integrated software stack with PyTorch, an open source software library for machine learning, and the Python programming language. The stack is designed for short and long-running high-performance tasks and is optimized for running on NVidia GPU.

Constructor:
Machine Learning
Pytorch
pytorch
0.4.0
Python
python
2.7.14
Cuda
cuda
9.1.85.3
Cudnn
cudnn
7.1.3
Nvidia drivers
cuda_only-nvidia_drivers
396.24
Preset
development_preset
1
CloudsInstallContainersDownloadConfiguration

Usage

You can install the appliance on any new or existing Linux server, download and run it as a virtual machine, use it as a base image for Docker or Vagrant, or launch it with a new cloud platform instance, VPS or dedicated server for supported hosting providers.

Launch in clouds
Aws ec2 Google cloud platform Digital ocean Microsoft azure Ibm bluemix Oracle cloud platform
Install
Install on Linux

You can install the appliance directly on any Linux with 64-bit kernel (>=2.6.32). Run from the command line:

curl -L http://jetware.io/appliances/aise/pytorch04_python2_cuda91-180512/file/installer:nub_tgz/setup | sh

You’ll be asked to execute some operations as root via sudo during the installation.

How to use

To enter the runtime environment or to execute a command inside the runtime environment you can use the utility /jet/enter. If no arguments are present, the standard shell will be executed inside the runtime environment. You can specify a command as an argument, it will be executed inside the runtime environment.

For example, to start all services in the runtime environment you can do /jet/enter start. To execute a mysql client you can do /jet/enter mysql; or run first /jet/enter, and than run from the new command line mysql.

Download archive

You can download the archive, unpack it into the /jet directory, finish installation by executing the command /jet/enter /jet/own/bin/fasten and start the services by running /jet/enter start.

machine_learning-tdhcd4tl8p15.tar.gz
2.35 GB
Download
Run a container
Alpine 3.8
Ubuntu 18.04
Debian 9
CentOS 7
Docker

Copy Dockerfile into your project’s directory:

FROM alpine:3.8
RUN apk add sudo libcap tar
RUN adduser -S -h /jet -s /jet/login jet

RUN { wget -O - "http://jetware.io/appliances/aise/pytorch04_python2_cuda91-180512/file/container:docker/setup" || echo exit 1; } | sh

WORKDIR /jet
USER jet
VOLUME ["/jet/prs"]
ENTRYPOINT ["/jet/login"]
CMD ["main"]

or build an image directly from the URL by executing the command:

docker build 'http://jetware.io/appliances/aise/pytorch04_python2_cuda91-180512/file/container:docker/Dockerfile?os=alpine_3.8'

Copy Dockerfile into your project’s directory:

FROM ubuntu:18.04
RUN \
  echo "tzdata tzdata/Areas select Etc" | debconf-set-selections; \
  echo "tzdata tzdata/Zones/Etc select Universal" | debconf-set-selections; \
  apt-get update && \
    DEBIAN_FRONTEND=noninteractive DEBCONF_NONINTERACTIVE_SEEN=true \
      apt-get install -y tzdata netbase sudo wget libcap2-bin
RUN useradd --system -m --home /jet --shell /jet/login jet

RUN { wget -O - "http://jetware.io/appliances/aise/pytorch04_python2_cuda91-180512/file/container:docker/setup" || echo exit 1; } | sh

WORKDIR /jet
USER jet
VOLUME ["/jet/prs"]
ENTRYPOINT ["/jet/login"]
CMD ["main"]

or build an image directly from the URL by executing the command:

docker build 'http://jetware.io/appliances/aise/pytorch04_python2_cuda91-180512/file/container:docker/Dockerfile?os=ubuntu_18.04'

Copy Dockerfile into your project’s directory:

FROM debian:9-slim
RUN apt-get update && apt-get install -y netbase sudo wget libcap2-bin
RUN useradd --system -m --home /jet --shell /jet/login jet

RUN { wget -O - "http://jetware.io/appliances/aise/pytorch04_python2_cuda91-180512/file/container:docker/setup" || echo exit 1; } | sh

WORKDIR /jet
USER jet
VOLUME ["/jet/prs"]
ENTRYPOINT ["/jet/login"]
CMD ["main"]

or build an image directly from the URL by executing the command:

docker build 'http://jetware.io/appliances/aise/pytorch04_python2_cuda91-180512/file/container:docker/Dockerfile?os=debian_9'

Copy Dockerfile into your project’s directory:

FROM centos:7
RUN yum install -y sudo wget libcap
RUN useradd --system -m --home /jet --shell /jet/login jet

RUN { wget -O - "http://jetware.io/appliances/aise/pytorch04_python2_cuda91-180512/file/container:docker/setup" || echo exit 1; } | sh

WORKDIR /jet
USER jet
VOLUME ["/jet/prs"]
ENTRYPOINT ["/jet/login"]
CMD ["main"]

or build an image directly from the URL by executing the command:

docker build 'http://jetware.io/appliances/aise/pytorch04_python2_cuda91-180512/file/container:docker/Dockerfile?os=centos_7'
Download a virtual machine
Ubuntu 14.04
VMWare
machine_learning-tdhcd4tl8p15-vmware-ubuntu_14.04.zip
Download
VirtualBox
machine_learning-tdhcd4tl8p15-virtualbox-ubuntu_14.04.ova
Download
Qemu/KVM
machine_learning-tdhcd4tl8p15-qemu-ubuntu_14.04.qcow2.gz
Download

You can access the virtual machine via console or SSH:

Login: jet
Password: jet
Configuration
Logo podval
© 2016— Jetware team