Although licenses are directly tied to training data and the resulting document type templates, there are a number of important points to note. Consider the following scenarios:
License Movement: When a product key (license) is moved from one server to another, the AI document types that were trained using that product key will be available on the new server. This means that the trained models themselves are associated with the license and can be transferred along with it.
Training Templates: However, the training templates, which are the foundational data and configurations used to train these AI document types, do not automatically transfer with the license. This is because the template data is stored locally on the original server and are not inherently linked to the product key.
Implications: Without the original training templates, the new server cannot perform further training or re-training of the AI document types. To enable this capability, the original server's data, including the training templates, would need to be backed up and restored on the new server. This ensures that all necessary components for training are available.
Training and Job Assignment: In this scenario, a document type is trained using a specific license and then assigned to a job. This job is backed up into a UMJ file, which is a backup format used by Umango to encapsulate job configurations and associated data.
Restoration by Another User: When this UMJ file is provided to another user with a different license, they can restore the job and execute it using the trained document type providing their AI server region is set to be the same as the region used to train the document type. This is because the UMJ file contains a reference to the necessary information to run the job with the trained model.
Limitations on New Jobs: However, the new user cannot create a new job and assign the same document type to it. This limitation arises because the source of the document type, which includes the training data and configurations, remains tied to the original license. The new user does not have access to the underlying training data or the ability to modify the document type beyond its current state.
License Dependency: The association of AI document types with the product key emphasizes the importance of the license in managing and utilizing trained models within Umango. It acts as a gatekeeper for the capabilities and distribution of these models.
Data Management: Proper data management, including backing up and transferring training templates and job configurations, is crucial for maintaining the flexibility and functionality of AI document types across different environments.
Operational Constraints: Users must be aware of the operational constraints imposed by the licensing model, particularly when it comes to transferring capabilities between servers or users. Understanding these constraints helps in planning and executing document processing workflows effectively.
On-Premise Installations: Users running fully on-premise licenses do not have the ability to use trained custom document types anywhere other than the server they were trained on. This is because the training data remains with the on-premise server and not in a cloud environment accessible to other servers.
Link to this article https://umango.com/KB?article=143