A deep learning model package 
(.dlpk) contains the files and
data required to run deep learning inferencing tools for object
detection or image classification. The package can be uploaded to
your portal as a DLPK item and used as the input to deep learning raster analysis
tools.
Deep learning model packages must contain an Esri model definition file (.emd) and a trained model file. The trained model file extension depends on the framework you used to train the model. For example, if you trained the model using TensorFlow, the model file will be a .pb file, while a model trained using Keras will generate an .h5 file. Depending on the model framework and options you used to train the model, you may need to include a Python raster function (.py) or additional files. You can include multiple trained model files in a single deep learning model package.
Create a deep learning model package
To create a deep learning model package, complete the following steps:
- In the Package group on the Share tab, click Deep Learning .
The Share As Deep Learning Package pane appears.
- Specify where to save the package, either to your online account or as a file on disk.
- Provide the name and, if saving the package to a file, the location for the new package on disk.
- Optionally, complete the Summary and Tags fields.
A summary and tags are required when sharing to an ArcGIS Enterprise 10.9 or earlier portal.
If you save the package to your online account, you can use the Tags drop-down menu to select existing tags. You can enter a maximum of 128 tags.
- If you choose to save the package to your online account, you can use the Categories drop-down menu to assign content categories.
You can assign a maximum of 20 categories.
- For Model Definition, provide the path to the .emd file.
- For Input, add the required items.
These can be files or folders, and must include the path to the trained model file (.pb, .h5, .pkl, and so on) at a minimum.
- Click Analyze to check for errors and issues.
Potential issues can include incorrect file paths, an invalid .emd file, and so on.
- Once the inputs are validated, click Package to create the deep learning model package.
Deep learning model packages can be saved locally or stored on your portal as a DLPK item. You can use the local .dlpk file as the input to deep learning tools in ArcGIS Pro. You can use the DLPK item in your portal to run the raster analysis deep learning tools in Map Viewer, ArcGIS API for Python, ArcGIS REST API, and ArcGIS Pro.