Vector 7.0 | Configuring Machine Learning Models (MLM)

System Administrator Guide > System Administrator Guide > Configuring Machine Learning Models (MLM)

Was this helpful?

Configuring Machine Learning Models (MLM)

MLM using Tensorflow

Tensorflow MLMs perform model scoring and inference in X100. Data does not move out of the engine which has several advantages:

• Query integration: Subsequent operations can be performed in the database engine.

• Reduced data transfer: Only (potentially smaller) query results are shipped.

• Scalability: ML inference can be a pipelined operation, so there is no need to materialize the full dataset. This would be needed when pulling data into a dedicated environment, or manual buffering strategies.

• Integrated and convenient usage.

Note: We only support saved model Tensorflow MLMs. Currently, h5 and Keras formats are not supported.

Last modified date: 12/19/2024