Available with Image Server
Review the browser and account requirements for and configure Deep Learning Studio.
Browser requirements
Use Deep Learning Studio in desktop web browsers. The app is not supported on mobile devices or tablets. The app is supported on the browsers listed below. For the best performance, use the latest version of the applicable browser.
- Google Chrome
- Microsoft Edge
- Mozilla Firefox
- Safari
ArcGIS Enterprise account requirements
To use Deep Learning Studio, your ArcGIS Enterprise account must meet the following requirements:
- For the analyst role in Deep Learning Studio and to access its full capabilities, you must be assigned the ArcGIS Enterprise Publisher or Administrator role, or a custom role with the following privileges:
- Create, update, and delete (content).
- Publish server-based layers.
- Use raster analysis tools.
- For the sample collector role in Deep Learning Studio, you must be assigned the ArcGIS Enterprise Editor user type, or a custom role with the following privileges:
- Create, update, and delete (content).
Supported languages
Deep Learning Studio supports all languages that ArcGIS Enterprise supports. The language used by the app is determined by the language settings in your account profile.
Configure Deep Learning Studio
Deep Learning Studio provides deep learning capabilities using distributed processing and storage in ArcGIS Enterprise. Follow the steps below to configure Deep Learning Studio in an ArcGIS Enterprise deployment.
- The ArcGIS Enterprise portal must be configured for raster analytics. This requires a federated ArcGIS Server to be licensed with the ArcGIS Image Server role and set as the raster analysis server. See Configure ArcGIS Server for deep learning raster analytics for details.
Note:
A file share raster store must be registered with the federated ArcGIS Image Server.
- Deep Learning Studio requires deep learning packages; see the Deep Learning Libraries Installation Guide for ArcGIS Image Server for instructions.
Note:
Both the CPU and GPU in your raster analysis server can be used for deep learning processing. If you plan to use a GPU in your environment, a NVIDIA GPU with CUDA Compute Capability is required.
- To see processing details for the geoprocessing processes, the administrator should set ArcGIS Image Server logging to the level of info. Set the Message Level value for the following geoprocessing services to INFO: RasterAnalysisTools.
- Deep Learning Studio supports image services published from 10.7 and later. Image services published from earlier versions are not supported.
- Verify that Deep Learning Studio is properly configured by confirming the app appears in the Apps Launcher.
- Configure and deploy ArcGIS Enterprise for raster analytics.
A file share type of raster store is required.
- Sign in to ArcGIS Enterprise to open the home page of your ArcGIS organization.
- Click the Apps Launcher button
and click the Deep Learning Studio button to start the Deep Learning Studio app.
For the Deep Learning Studio app to automatically appear in the Apps Launcher, your account must have editing privileges in ArcGIS Enterprise. To access all the capabilities of Deep Learning Studio, you must have the appropriate privileges for content creation and service publishing.