mirror of
https://github.com/saymrwulf/onnxruntime.git
synced 2026-06-25 02:50:42 +00:00
Docker containers for CPU and GPU quickstart (#332)
* Docker Container for CPU Version (Ubuntu 16.04, Python3 Bindings, compatible with Windows-Docker) * Nvidia-Docker Container for GPU Version (Ubuntu 16.04, CUDA, CUDNN, Python3 Bindings) * README with Docker quickstart instructions (i.e. docker pull .../onnxruntime:cpu, docker run -it ...) * Include plans to publish public images (with ONNX Runtime 0.2) on README
This commit is contained in:
Vinitra Swamy
GitHub
parent
583631adf5
commit
ea94cbf14b
3 changed files with 115 additions and 0 deletions
23
dockerfiles/Dockerfile.cpu
Normal file
23
dockerfiles/Dockerfile.cpu
Normal file
|
|
@ -0,0 +1,23 @@
|
|||
#-------------------------------------------------------------------------
|
||||
# Copyright (c) Microsoft Corporation. All rights reserved.
|
||||
# Licensed under the MIT License.
|
||||
#--------------------------------------------------------------------------
|
||||
# Official user quickstart docker container for ONNX Runtime
|
||||
# Ubuntu 16.04, CPU version, Python 3.
|
||||
#--------------------------------------------------------------------------
|
||||
|
||||
FROM ubuntu:16.04
|
||||
MAINTAINER Vinitra Swamy "viswamy@microsoft.com"
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y sudo \
|
||||
build-essential curl \
|
||||
libcurl4-openssl-dev \
|
||||
libssl-dev wget \
|
||||
python 3.6 python3-pip \
|
||||
python3-dev git
|
||||
RUN pip3 install --upgrade pip
|
||||
RUN pip3 install numpy onnx
|
||||
|
||||
RUN pip3 install onnxruntime
|
||||
WORKDIR /code
|
||||
23
dockerfiles/Dockerfile.gpu
Normal file
23
dockerfiles/Dockerfile.gpu
Normal file
|
|
@ -0,0 +1,23 @@
|
|||
#-------------------------------------------------------------------------
|
||||
# Copyright (c) Microsoft Corporation. All rights reserved.
|
||||
# Licensed under the MIT License.
|
||||
#--------------------------------------------------------------------------
|
||||
# Official user quickstart nvidia-docker container for ONNX Runtime GPU
|
||||
# Ubuntu 16.04, GPU version, CuDNN 7, CUDA 10, Python 3.
|
||||
#--------------------------------------------------------------------------
|
||||
|
||||
FROM nvidia/cuda:10.0-cudnn7-devel-ubuntu16.04
|
||||
MAINTAINER Vinitra Swamy "viswamy@microsoft.com"
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y sudo \
|
||||
build-essential curl \
|
||||
libcurl4-openssl-dev \
|
||||
libssl-dev wget \
|
||||
python 3.6 python3-pip \
|
||||
python3-dev git
|
||||
RUN pip3 install --upgrade pip
|
||||
RUN pip3 install numpy onnx
|
||||
|
||||
RUN pip3 install onnxruntime-gpu
|
||||
WORKDIR /code
|
||||
69
dockerfiles/README.md
Normal file
69
dockerfiles/README.md
Normal file
|
|
@ -0,0 +1,69 @@
|
|||
# Quick-start Docker containers for ONNX Runtime
|
||||
|
||||
## CPU Version (Preview)
|
||||
#### Linux 16.04, Python Bindings, Compatible with Docker for Windows
|
||||
|
||||
1. Retrieve your docker image in one of the following ways.
|
||||
|
||||
- Build the docker image from the DockerFile in this repository.
|
||||
```
|
||||
# If you have a Linux machine, preface this command with "sudo"
|
||||
docker build -t onnxruntime-cpu -f Dockerfile.cpu .
|
||||
```
|
||||
- Pull the official image from DockerHub.
|
||||
|
||||
```
|
||||
# Will be available with ONNX Runtime 0.2.0
|
||||
```
|
||||
2. Run the docker image
|
||||
|
||||
```
|
||||
# If you have a Linux machine, preface this command with "sudo"
|
||||
# If you have a Windows machine, preface this command with "winpty"
|
||||
|
||||
docker run -it onnxruntime-cpu
|
||||
```
|
||||
|
||||
## GPU Version (Preview)
|
||||
#### Linux 16.04, Python Bindings, CUDA 10, CuDNN7, Requires Nvidia-Docker version 2.0
|
||||
|
||||
0. Prerequisites: [Install Nvidia-Docker 2.0](https://github.com/nvidia/nvidia-docker/wiki/Installation-(version-2.0))
|
||||
|
||||
1. Retrieve your docker image in one of the following ways.
|
||||
- Build the docker image from the DockerFile in this repository.
|
||||
```
|
||||
# If you have a Linux machine, preface this command with "sudo"
|
||||
|
||||
docker build -t onnxruntime-gpu -f Dockerfile.gpu .
|
||||
```
|
||||
Note that you can change the base CUDA distribution to 9.1 and use nvidia-docker v1
|
||||
by replacing the first line of the dockerfile with the base image below.
|
||||
```
|
||||
FROM nvidia/cuda:9.1-cudnn7-devel-ubuntu16.04
|
||||
```
|
||||
- Pull the official image from DockerHub.
|
||||
|
||||
```
|
||||
# Will be available with ONNX Runtime 0.2.0
|
||||
```
|
||||
|
||||
2. Run the docker image
|
||||
```
|
||||
# If you have a Linux machine, preface this command with "sudo"
|
||||
# If you have a Windows machine, preface this command with "winpty"
|
||||
|
||||
docker run -it --runtime=nvidia --rm nvidia/cuda onnxruntime-gpu
|
||||
```
|
||||
### Other options to get started with ONNX Runtime
|
||||
|
||||
- Deploy [inference for pretrained ONNX models](https://github.com/Azure/MachineLearningNotebooks/tree/master/how-to-use-azureml/deployment/onnx) for handwritten digit recognition (MNIST)
|
||||
or facial expression recognition (FER+) using Azure Machine Learning
|
||||
|
||||
- Work with ONNX runtime in your local environment using the PyPi release ([CPU](https://pypi.org/project/onnxruntime/), [GPU](https://pypi.org/project/onnxruntime-gpu/))
|
||||
- ``pip install onnxruntime``
|
||||
- ``pip install onnxruntime-gpu``
|
||||
|
||||
- Build ONNX Runtime from the source code by following [these instructions for developers](../BUILD.md).
|
||||
|
||||
### License
|
||||
[MIT License](../LICENSE)
|
||||
Loading…
Reference in a new issue