Implementation:Interpretml Interpret UNTESTED From Jsonable

Field	Value
Sources	Repo: InterpretML
Domains	Model_Persistence, Interoperability
Updated	2026-02-07

Overview

Concrete tool for deserializing a JSON dictionary back into a fitted EBM model provided by the InterpretML library (experimental).

Description

The UNTESTED_from_jsonable function populates an empty EBM model instance from a JSON dictionary. It restores all fitted attributes (bins, scores, intercept, feature metadata), converts types back to numpy arrays, and sets has_fitted_ = True. The UNTESTED prefix indicates this function is considered experimental.

Usage

Use to load a previously saved EBM model from JSON for inference or further analysis.

Code Reference

Field	Value
Source	interpretml/interpret
File	python/interpret-core/interpret/glassbox/_ebm/_json.py
Lines	415-627

Signature:

def UNTESTED_from_jsonable(ebm, jsonable):
    """Converts JSON into a model.
    Args:
        jsonable: the JSONable object
    Returns:
        An EBM
    """

Import:

from interpret.glassbox._ebm._json import UNTESTED_from_jsonable

I/O Contract

Inputs:

Parameter	Type	Required	Description
ebm	empty EBM instance	Yes	An uninitialized ExplainableBoostingClassifier or ExplainableBoostingRegressor
jsonable	dict	Yes	Dictionary with "ebm" key containing the serialized model state

Outputs:

Type	Description
None	Modifies ebm in-place, sets has_fitted_ = True

Usage Examples

import json
from interpret.glassbox import ExplainableBoostingClassifier
from interpret.glassbox._ebm._json import UNTESTED_from_jsonable

# Load from file
with open("ebm_model.json", "r") as f:
    model_dict = json.load(f)

ebm = ExplainableBoostingClassifier()
UNTESTED_from_jsonable(ebm, model_dict)
# ebm is now ready for predictions
predictions = ebm.predict(X_test)

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment