What's new in Maximo Visual Inspection 8.7

Data migration

When you upgrade from Maximo Visual Inspection 8.6.1 to 8.7, your data is migrated. Ensure that you take the following precautions before you upgrade:

• Protect your data. The existing MongoDB pod and PVC remain. You can access the old data for manual removal.
• Stop all inspections. Consistency changes can occur if inspections are not stopped before the upgrade.

Out-of-domain (OOD) detection

You can enable OOD detection for YOLO v3 models that use the CoreML training option. You enable OOD detection when you train the YOLO v3 models. After the model is trained and configured in an inspection in IBM Maximo Visual Inspection Mobile, you can export inspection result data sets from Maximo Visual Inspection or by using the REST API. The results contain an OOD score. The OOD score indicates how similar the image is to the images that are in the data set that was used to train the model. When OOD detection is enabled, the inference is slower.

Online augmentation for high-resolution model training

You can enable online augmentation when you are training high-resolution models. A sequence of four probabilistic transformation types is applied to each image in the data set during model training. The transformation types increase a model's ability to generalize so that you do not need to apply offline augmentation to a data set before training.

Upgraded NVIDIA support

You can train and deploy high-resolution, anomaly optimized, and action detection models on NVIDIA Ampere GPUs. You can also combine this model type with other models in the same container and deploy the models locally on Maximo Visual Inspection, subject to resource limitations. Train and deploy action detection models on Ampere cards.

MongoDB

Maximo Visual Inspection now supports the use of the IBM Maximo Application Suite MongoDB instance without the need for configuration changes.

Grafana dashboards

You can create and access a Grafana dashboard. In the dashboard, you can visualize runtime metrics, such as training, and inference metrics. For more information about Grafana dashboards, see Installing Grafana.

Anomaly models and GPU memory

Anomaly models that were trained in earlier versions of Maximo Visual Inspection can take up too much GPU memory when the models are deployed in Maximo Visual Inspection 8.7. This situation is affected by the number of images that trained the model and the available memory on the corresponding GPU. Anomaly models that are trained in Maximo Visual Inspection 8.7 require up to 8 GB of memory on the corresponding GPU.

Documentation improvements

To optimize and improve content retrievability, the documentation is now organized and streamlined to identify information by product and version. For more information, see Documentation conventions.

Data sets tour

You can take a tour of the Data sets page to learn about the page and how to manage data sets. To get started, click Take a tour.

End of support for PostgreSQL

Starting in Maximo Visual Inspection 8.7, postgreSQL and keycloak are no longer supported to provide minimal exposure to security risks. PostgreSQL and keycloak are still supported in Maximo Visual Inspection 8.6 and earlier versions.

End of support for custom models and custom scripts

Starting in Maximo Visual Inspection 8.7, custom models are not supported and custom scripts cannot be uploaded. Custom models are still supported in Maximo Visual Inspection 8.6 and earlier versions. REST APIs still support the retrieval of any existing custom scripts that you uploaded to Maximo Visual Inspection 8.6 and earlier versions.

For more information about custom models and custom scripts in a supported version, see Working with custom models in Maximo Visual Inspection version 8.5 or the previous supported versions.