Fusion 5.12
Product Selector
Use a TensorFlow deep learning model to encode and clusterize queries into a fixed dense vector space.
This stage is used in the default query pipeline that enables Smart Answers. See Detailed Pipeline Setup for configuration guidelines.
| If you are using a dynamic schema, make sure this stage is added after the Solr Dynamic Field Name Mapping stage. |
Use a deep learning model to encode and clusterize queries into fixed dense vector space
Set to true to skip this stage.
Default: false
A unique label for this stage.
<= 255 characters
Define a conditional script that must result in true or false. This can be used to determine if the stage should process or not.
The ID of the DL encoder model bundled to the MLeap format and stored in the Fusion blob store.
>= 1 characters
Name of the document field to feed into the encoder.
>= 1 characters
Default: body_t
Name of the field to store the encoded vector in the document. "_ds" suffix will be added if not already present
>= 1 characters
Default: document_vector_ds
Name of the field to store document extracted clusters. "_ss" suffix will be added if not already present
>= 1 characters
Default: document_clusters_ss
Name of the field to store document extracted distances to the clusters. "_ds" suffix will be added if not already present
>= 1 characters
Default: document_distances_ds
Number of document clusters to be stored in clusters field. Should be less or equal to the amount of clusters from the encoder model.
Default: 1
Flag to indicate if this stage should throw an exception if an error occurs.
Default: false
Flag to indicate that the encoded document vector and clusters should be stored in the Context instead of the documents fields.
Default: false