Product Selector

Fusion 5.9
    Fusion 5.9

    Train a New ModelLucidworks AI

    This feature is currently only available to clients who have contracted with Lucidworks for features related to Neural Hybrid Search and Lucidworks AI.

    This topic details how to train a new Lucidworks AI custom model using the Lucidworks AI custom embedding model user interface.

    For technical information about parameters and configuration for custom embedding models, see Custom embedding model training.

    To train a new custom model, complete the following:

    1. In the Lucidworks AI screen, click the Custom Models tab.

    2. In the Custom Models screen, click + Train New Model.

    3. In the Train New Model screen, review the information about how to generate your service account key. To view an example, click Sample service account key.

    4. In the Model Type field, select either General, which is typically used by informational sites or eCommerce for online sales sites.

    5. In the Signals Training Data field, enter the location of signals in the training data in Google Cloud Storage (GCS). For configuration information, see Training data query file.

    6. In the Catalog Training Data field, enter the location of the catalog of the training data in Google Cloud Storage (GCS). For configuration information, see Training data index file.

    7. In the Upload Service Account Key section, click Upload.

      If your Google Cloud Storage (GCS) bucket is publicly accessible, the service account key is optional and you can click Continue without entering a key. If you enter a key, it is validated when you click Continue. If the key is not valid, the model cannot be configured.

    8. Click Continue to configure the model.

    Basic model configuration

    1. In the Name field, enter the name of the new model.

    2. In the Region field, select the region where the model is deployed.

      This selection cannot be changed. If you need to train a model in a different region, you need to create another model, specify that other region, and then train the new model.

    3. Click Continue.

      The Guided Entry screen is displayed. If you want to enter a configuration, click Switch to Manual Entry.

    Guided entry

    1. In Layer 1 of the RNN Layers field:

      • Select the bi-directional recurrent neural network (RNN) layers to use. Options are gru and lstm.

      • Select the unit size of the layer.

        Layer 2 is optional. If needed, select the RNN and unit.

    2. In the Text Processor Config field, select the type of tokenization and embedding to be used as the base for the selected RNN layer. Two primary options are word or byte pair encoding (BPE).

    3. Click Save & Run to execute the model training. To exit without saving and running the model, click Cancel.

    Manual entry

    If you click Switch to Manual Entry, the Custom Config field displays.

    1. In the Custom Config field, enter the configuration parameters for your custom model.

      For detailed information about configuration options for a model, see Custom configuration.

    2. Click Save & Run to execute the model training. To exit without saving and running the model, click Cancel. To return to creating a model without entering custom configuration, click Switch to Guided Entry.