API Documentation

API Documentation#

Ensemble Graph Constructors#

Ensemble(ops, schema[, label_columns])

Class that represents an entire ensemble consisting of multiple models that run sequentially in tritonserver initiated by an inference request.

Ensemble Operator Constructors#

workflow.TransformWorkflow([workflow, ...])

This operator takes a workflow and turns it into a ensemble operator so that we can execute feature engineering during ensemble on tritonserver.

tensorflow.PredictTensorflow(model_or_path)

TensorFlow Model Prediction Operator.

fil.PredictForest(model, input_schema, *[, ...])

Operator for running inference on Forest models.

implicit.PredictImplicit(model[, ...])

Operator for running inference on Implicit models..

softmax_sampling.SoftmaxSampling(relevance_col)

Given inputs of ID and prediction, this operator will sort all inputs in descending order.

session_filter.FilterCandidates(filter_out)

This operator takes the input column and filters out elements of that column based on the supplied criteria.

unroll_features.UnrollFeatures(item_id_col, ...)

This operator takes a target column and joins the "unroll" columns to the target.

Conversion Functions for Triton Inference Server#

convert_df_to_triton_input(schema, batch[, ...])

Convert a dataframe to a set of Triton inputs

convert_triton_output_to_df(columns, response)

Convert a Triton response to a dataframe
Contents