Nota
L'accesso a questa pagina richiede l'autorizzazione. È possibile provare ad accedere o modificare le directory.
L'accesso a questa pagina richiede l'autorizzazione. È possibile provare a modificare le directory.
È possibile usare la Data Science Virtual Machine (DSVM) per esplorare facilmente i dati e gestire l'apprendimento automatico nel cloud. Una DSVM è preconfigurata con patch di sicurezza, driver, software di data science e sviluppo diffusi e un sistema operativo completo. È possibile scegliere l'ambiente hardware adatto per l'utente, da computer incentrati sulla CPU a costi inferiori a computer potenti con più GPU, archiviazione NVMe e grandi quantità di memoria. Per i computer con GPU, vengono installati tutti i driver e tutti i framework di Machine Learning corrispondono alla versione per la compatibilità con GPU. Inoltre, l'accelerazione è abilitata in tutti i software dell'applicazione che supportano GPU.
La DSVM include gli strumenti di data science più utili preinstallati.
Creare soluzioni di Deep Learning e Machine Learning
| Strumento | DSVM Windows Server 2019 | Windows Server 2022 DSVM | Ubuntu 20.04 DSVM | Note sull'utilizzo |
|---|---|---|---|---|
| Driver CUDA, cuDNN, NVIDIA | ✅ | ✅ | ✅ |
CUDA, cuDNN, NVIDIA Driver nella DSVM |
| Horovod | ❌ | ❌ | ✅ | Horovod nella DSVM |
| NVidia System Management Interface (nvidia-smi) | ✅ | ✅ | ✅ | nvidia-smi nella DSVM |
| PyTorch | ✅ | ✅ | ✅ | PyTorch nella DSVM |
| TensorFlow | ✅ | ✅ |
✅ | TensorFlow nella DSVM |
| Integrazione con Azure Machine Learning (Python) | ✅ (Python SDK, esempi) |
✅ (Python SDK, esempi) |
✅ (Python SDK, interfaccia della riga di comando, esempi) |
SDK di Azure Machine Learning |
| XGBoost | ✅ (supporto CUDA) |
✅ (supporto CUDA) |
✅ (supporto CUDA) |
XGBoost nella DSVM |
| Vowpal Wabbit | ✅ | ✅ | ✅ |
Vowpal Wabbit nella DSVM |
| Weka | ❌ | ❌ | ❌ | |
| LightGBM | ❌ | ❌ | ✅ (GPU, supporto MPI) |
|
| H2O | ❌ | ❌ | ✅ | |
| CatBoost | ❌ | ❌ | ✅ | |
| Intel MKL | ❌ | ❌ | ✅ | |
| OpenCV | ❌ | ❌ | ✅ | |
| Dlib | ❌ | ❌ | ✅ | |
| Docker | ✅ (solo contenitori di Windows) |
✅ (solo contenitori di Windows) |
✅ | |
| Nccl | ❌ | ❌ | ✅ | |
| Rattle | ❌ | ❌ | ❌ | |
| PostgreSQL | ❌ | ❌ | ✅ | |
| Runtime ONNX | ❌ | ❌ | ✅ |
Archiviare, recuperare e modificare i dati
| Strumento | DSVM Windows Server 2019 | Windows Server 2022 DSVM | Ubuntu 20.04 DSVM | Note sull'utilizzo |
|---|---|---|---|---|
| Database relazionali | SQL Server 2019 Developer Edition |
SQL Server 2019 Developer Edition |
SQL Server 2019 Developer Edition |
SQL Server nella DSVM |
| Strumenti del database | SQL Server Management Studio SQL Server Integration Services bcp, sqlcmd |
SQL Server Management Studio SQL Server Integration Services bcp, sqlcmd |
SQuirreL SQL (strumento di query), bcp, sqlcmd Driver ODBC/JDBC |
|
| Azure Storage Explorer | ✅ |
✅ |
||
| Interfaccia della riga di comando di Azure | ✅ |
✅ |
✅ |
|
| AzCopy | ✅ |
✅ |
❌ | AzCopy nella DSVM |
| Driver BLOB FUSE | ❌ | ❌ | ❌ |
blobfuse nella DSVM |
| Utilità di migrazione dati di Azure Cosmos DB | ✅ | ✅ | ❌ | Azure Cosmos DB nella DSVM |
| Strumenti da riga di comando Unix/Linux | ❌ | ❌ | ✅ | |
| Apache Spark 3.1 (autonomo) | ✅ | ✅ | ✅ |
Programmare in Python, R, Julia e Node.js
| Strumento | DSVM Windows Server 2019 | Windows Server 2022 DSVM | Ubuntu 20.04 DSVM | Note sull'utilizzo |
|---|---|---|---|---|
| CRAN-R con i pacchetti più diffusi preinstallati | ✅ | ✅ | ✅ | |
| Anaconda Python con i pacchetti più diffusi preinstallati | ✅ | ✅ (Miniconda) |
✅ (Miniconda) |
|
| Julia (Julialang) | ✅ | ✅ | ✅ | |
| JupyterHub (server notebook multiutente) | ❌ | ❌ | ✅ | |
| JupyterLab (server notebook multiutente) | ✅ | ✅ | ✅ | |
| Node.js | ✅ | ✅ | ✅ | |
| Server Jupyter Notebook con i kernel seguenti: | ✅ |
✅ |
✅ | Esempi di Jupyter Notebook |
| R | Esempi di R Jupyter | |||
| Python | Esempi di Python Jupyter | |||
| Julia | Esempi di Julia Jupyter | |||
| PySpark | Esempi di PySpark Jupyter |
Ubuntu 20.04 DSVM, Windows Server 2019 DSVM e Windows Server 2022 DSVM hanno questi kernel Jupyter:
- Python3.8-default
- Python3.8-Tensorflow-Pytorch
- Python3.8-AzureML
- R
- Python 3.7 - Spark (locale)
- Julia 1.6.0
- R Spark - HDInsight
- Scala Spark - HDInsight
- Python 3 Spark - HDInsight
Ubuntu 20.04 DSVM, Windows Server 2019 DSVM e Windows Server 2022 DSVM hanno gli ambienti conda seguenti:
- Python3.8-default
- Python3.8-Tensorflow-Pytorch
- Python3.8-AzureML
Usare l'editor o l'IDE preferito
| Strumento | DSVM Windows Server 2019 | Windows Server 2022 DSVM | Ubuntu 20.04 DSVM | Note sull'utilizzo |
|---|---|---|---|---|
| Notepad++ | ✅ |
✅ |
❌ |
|
| Nano | ✅ |
✅ |
❌ |
|
| Visual Studio 2019 Community Edition | ✅ |
✅ | ❌ | Visual Studio nella DSVM |
| Visual Studio Code | ✅ |
✅ |
✅ |
Visual Studio Code nella DSVM |
| PyCharm Community Edition | ✅ |
✅ |
✅ |
PyCharm nella DSVM |
| IntelliJ IDEA | ❌ | ❌ | ✅ | |
| Vim | ❌ | ❌ | ✅ |
|
| Emacs | ❌ | ❌ | ✅ |
|
| Git e Git Bash | ✅ |
✅ |
✅ |
|
| OpenJDK 11 | ✅ |
✅ |
✅ |
|
| .NET Framework | ✅ |
✅ |
❌ | |
| Azure SDK | ✅ |
✅ | ✅ |
Organizzare e presentare i risultati
| Strumento | DSVM Windows Server 2019 | Windows Server 2022 DSVM | Ubuntu 20.04 DSVM | Note sull'utilizzo |
|---|---|---|---|---|
| Microsoft 365 (Word, Excel, PowerPoint) | ✅ | ✅ | ❌ | |
| Microsoft Teams | ✅ | ✅ | ❌ | |
| Power BI Desktop | ✅ | ✅ |
❌ | |
| Browser Microsoft Edge | ✅ | ✅ | ✅ |