This CentOS 7.4-based data science virtual machine (DSVM) contains popular tools for data science and development activities, including Microsoft R Open, Anaconda Python, Azure command line tools, and xgboost. It also contains deep learning components: the NVIDIA driver, CUDA, cuDNN, TensorFlow, Microsoft Cognitive Toolkit, PyTorch, and several other frameworks.
What's new
The CentOS DSVM now supports deep learning on Azure GPU instances (the NC, NCv2, and ND series) with the NVIDIA drivers, CUDA, cuDNN, and GPU versions of TensorFlow (with Keras), Microsoft Cognitive Toolkit, MXNet, Chainer, and PyTorch.
The CentOS DSVM also now includes Microsoft ML Server 9.2.1, with additional support for operationalizing R models, Python machine learning modules, pre-trained models, and many more features.
Other Editions
We also offer the DSVM on Windows 2016 and Ubuntu and recommend that users choose one of those versions if possible.
Included in the CentOS DSVM
- Microsoft Microsoft ML Server 9.2.1, with Microsoft R Open 3.4.1, the MicrosoftML package for Python, new operationalization features, and many more
- Anaconda Python 5 with Conda environments for 2.7 and 3.5
- Jupyter support via JupyterHub with many sample notebooks ready to run
- Deep learning with TensorFlow, Keras, Microsoft Cognitive Toolkit, MXNet, Chainer, and PyTorch, as well as CUDA 8 and cuDNN 6
- Spark local 2.2.0 with a PySpark Jupyter kernel
- Single node local Hadoop with HDFS and Yarn
- The Azure CLI, Azure Storage Explorer, several SDKs, the Azure ML Model Management CLI, and the Azure Blob storage FUSE library
- Docker and NVIDIA Docker
- xgboost (with CUDA support)
- Vowpal Wabbit for online learning
- Apache Drill for querying non-relational data using SQL
- Visual Studio Code IDEs, IntelliJ IDEA, PyCharm, and Atom
- JuliaPro, a curated distribution of Julia Language and tools
You can view a full list of installed tools for the Linux edition here.
