Delete an ML Model

Delete an ML model by ID or name, with soft/hard delete options

DELETE /v1/mlmodels/{id}

DELETE /v1/mlmodels/{id}

Python

from metadata.sdk import configure
from metadata.sdk.entities import MLModels

configure(
    host="https://your-company.open-metadata.org/api",
    jwt_token="your-jwt-token"
)

# Soft delete by ID
MLModels.delete("6b04e1d8-b66d-4f78-ab21-beb5be2cf4f2")

# Hard delete
MLModels.delete(
    "6b04e1d8-b66d-4f78-ab21-beb5be2cf4f2",
    hard_delete=True
)

# Delete by name
MLModels.delete_by_name("mlflow_svc.customer_segmentation")

# Delete by name with options
MLModels.delete_by_name(
    "mlflow_svc.customer_segmentation",
    hard_delete=True
)

# Restore a soft-deleted ML model
MLModels.restore("6b04e1d8-b66d-4f78-ab21-beb5be2cf4f2")

Response (Soft Delete)

{
  "id": "6b04e1d8-b66d-4f78-ab21-beb5be2cf4f2",
  "name": "customer_segmentation",
  "fullyQualifiedName": "mlflow_svc.customer_segmentation",
  "displayName": "Customer Segmentation Model",
  "algorithm": "KMeans",
  "version": 0.2,
  "updatedAt": 1769982669247,
  "updatedBy": "admin",
  "service": {
    "id": "ca22d46e-81b9-4e48-85b5-0adc44980da9",
    "type": "mlmodelService",
    "name": "mlflow_svc",
    "fullyQualifiedName": "mlflow_svc"
  },
  "serviceType": "Mlflow",
  "deleted": true,
  "owners": [],
  "tags": [],
  "domains": []
}

Delete an ML Model

Delete an ML model by ID or fully qualified name. Supports soft delete (default), hard delete, and restore operations.

Delete by ID

id · string · required

UUID of the ML model to delete.

recursive · boolean · default: "false"

Recursively delete child entities.

hardDelete · boolean · default: "false"

Permanently delete the ML model. If false, the ML model is soft-deleted and can be restored.

Delete by Name

Use DELETE /v1/mlmodels/name/{fqn} to delete by fully qualified name.

fqn · string · required

Fully qualified name of the ML model (e.g., mlflow_svc.customer_segmentation).

recursive · boolean · default: "false"

Recursively delete child entities.

hardDelete · boolean · default: "false"

Permanently delete the ML model.

Restore a Soft-Deleted ML Model

Use PUT /v1/mlmodels/restore to restore a soft-deleted ML model.

id · string · required

UUID of the soft-deleted ML model to restore.

DELETE /v1/mlmodels/{id}

Python

from metadata.sdk import configure
from metadata.sdk.entities import MLModels

configure(
    host="https://your-company.open-metadata.org/api",
    jwt_token="your-jwt-token"
)

# Soft delete by ID
MLModels.delete("6b04e1d8-b66d-4f78-ab21-beb5be2cf4f2")

# Hard delete
MLModels.delete(
    "6b04e1d8-b66d-4f78-ab21-beb5be2cf4f2",
    hard_delete=True
)

# Delete by name
MLModels.delete_by_name("mlflow_svc.customer_segmentation")

# Delete by name with options
MLModels.delete_by_name(
    "mlflow_svc.customer_segmentation",
    hard_delete=True
)

# Restore a soft-deleted ML model
MLModels.restore("6b04e1d8-b66d-4f78-ab21-beb5be2cf4f2")

Response (Soft Delete)

{
  "id": "6b04e1d8-b66d-4f78-ab21-beb5be2cf4f2",
  "name": "customer_segmentation",
  "fullyQualifiedName": "mlflow_svc.customer_segmentation",
  "displayName": "Customer Segmentation Model",
  "algorithm": "KMeans",
  "version": 0.2,
  "updatedAt": 1769982669247,
  "updatedBy": "admin",
  "service": {
    "id": "ca22d46e-81b9-4e48-85b5-0adc44980da9",
    "type": "mlmodelService",
    "name": "mlflow_svc",
    "fullyQualifiedName": "mlflow_svc"
  },
  "serviceType": "Mlflow",
  "deleted": true,
  "owners": [],
  "tags": [],
  "domains": []
}

---

Returns

Soft delete returns the ML model object with deleted: true. Hard delete returns no content (204). Restore returns the restored ML model object.

---

Error Handling

Code	Error Type	Description
401	UNAUTHORIZED	Invalid or missing authentication token
403	FORBIDDEN	User lacks permission to delete this ML model
404	NOT_FOUND	ML model with given ID or FQN does not exist