.. _auto_ml_predict_interactive:

*ML PREDICT INFO*
=================

Runs interactive inference against a previously trained ML model, presenting a slider-based UI when input/target/schema metadata are supplied. Accepts a serialized model file (`.pkl` or per-target CSV manifest) and a user-defined prediction dataset, then returns a dataset of predicted values — including curve-shaped outputs assembled from indexed prediction columns.

When to use
-----------

Tagged: ``curve-prediction``, ``inference``, ``info``, ``interactive``, ``lucy-ml``, ``ml``, ``per-target``, ``pkl``.

Inputs
------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20 20 20

   * - Label
     - ID
     - Type
     - Default
     - Required
     - Description
   * - Model File
     - mfile
     - text
     - —
     - ✓
     - Path, attachment ID, or mathmodel ID pointing to the serialized model artifact — accepts a single `.pkl` file, a comma-separated list of per-target `.pkl` paths, or a CSV manifest produced by ML_PREDICT_INFO; required.
   * - Inputs
     - inputs
     - scalar
     - —
     - 
     - Comma-separated or structured list of input feature names as recorded during training; when wired from an ML_PREDICT_INFO worker it enables the interactive slider UI — leave empty if supplying a raw dataset directly.
   * - Targets
     - targets
     - scalar
     - —
     - 
     - Comma-separated or structured list of target/response names as recorded during training; when wired from an ML_PREDICT_INFO worker it enables the interactive slider UI — leave empty if not needed.
   * - Schema
     - schema
     - dataset
     - —
     - 
     - Dataset carrying the column schema (feature ranges, types, encoding metadata) captured at training time; when provided alongside `inputs` and `targets`, drives the slider UI for interactive prediction.
   * - Inputs For Prediction
     - dataset
     - predict_interactive
     - —
     - ✓
     - Tabular dataset of input rows for which predictions are requested; when `inputs`, `targets`, and `schema` are connected, an interactive slider is rendered to build this dataset — required.
   * - Reference Dataset
     - reference_dataset
     - dataset
     - —
     - 
     - Optional original training dataset that contains raw curve columns; used to overlay ground-truth curves against predictions — leave unset if curve comparison is not needed.
   * - Raw Column Name
     - raw_column_name
     - scalar
     - —
     - 
     - Column name in the training data that holds raw curve arrays; required only when the prediction target represents discrete curve points (e.g. `force_curve`) — leave empty for scalar targets.

Outputs
-------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20

   * - Label
     - ID
     - Type
     - Description
   * - Predictions
     - dataset
     - dataset
     - Tabular dataset of model predictions keyed by the original row IDs; scalar targets appear as named columns and curve targets are assembled into CurveGroup objects from indexed `<target>_<i>_pred` columns.

Disciplines
-----------

- ai_ml.supervised.classification
- ai_ml.supervised.regression
- ai_ml.surrogate

.. raw:: html

   <hr style="margin-top:2em">
   <p style="font-size:11px;color:#888">
   Auto-generated from <code>platform</code> schema. Worker id: <code>ml_predict_interactive</code>. Schema hash: <code>fc07283e6b2c</code>. Hand-curated docs in <code>workerexamples/</code> override this page when present.
   </p>