Update Model Configuration

Page Outline

• Path Parameters
• Request Body
• Behavior
• Response

---

Replaces or partially updates the training configuration for a model. Changes take effect on the next training run.

Base Path: /api/v1/models

Method: PUT

Path: /{model_id}/config

Auth: Requires a valid Bearer token (Authorization: Bearer <JWT>)

Path Parameters

Parameter	Type	Description
model_id	string	Unique identifier of the model

Request Body

{
  "algorithm": "lightgbm",
  "target_column": "churned",
  "feature_columns": ["age", "tenure", "monthly_spend", "support_tickets"],
  "hyperparameters": {
    "n_estimators": 300,
    "max_depth": 8,
    "learning_rate": 0.03
  },
  "preprocessing": {
    "handle_missing": "median",
    "scale_features": true
  }
}

Field	Type	Description
algorithm	string	Algorithm override: xgboost, lightgbm, random_forest, logistic_regression, auto
target_column	string	Column name to predict
feature_columns	string[]	List of input feature column names
hyperparameters	object	Key-value pairs passed directly to the algorithm
preprocessing	object	Preprocessing pipeline configuration

Behavior

• Uses PUT semantics — the entire config is replaced. Use Get Config first to ensure you preserve any fields you do not intend to change.
• Config updates are blocked while the model is in training status; returns HTTP 409.
• Validates that target_column and feature_columns do not overlap.
• An audit log entry is written with action model.config_updated, recording the previous and new config.

Response

Returns the full updated ModelConfig object (same schema as Get Config).