Logged Model

View as Markdown

A logged model message includes logged model attributes, tags, registration info, params, and linked run metrics.

GetGET /api/2.0/mlflow/logged-models/{model_id}CreatePOST /api/2.0/mlflow/logged-modelsDeleteDELETE /api/2.0/mlflow/logged-models/{model_id}Finalize Logged ModelPATCH /api/2.0/mlflow/logged-models/{model_id}Search Logged ModelsPOST /api/2.0/mlflow/logged-models/searchSet Logged Model TagsPATCH /api/2.0/mlflow/logged-models/{model_id}/tagsDelete Logged Model TagDELETE /api/2.0/mlflow/logged-models/{model_id}/tags/{tag_key}Log Logged Model ParamsPOST /api/2.0/mlflow/logged-models/{model_id}/params

LoggedModel object

A logged model message includes logged model attributes, tags, registration info, params, and linked run metrics.

infoobject

The logged model attributes such as model ID, status, tags, etc.

Show child attributesHide child attributes

model_idstring

The unique identifier for the logged model.

experiment_idstring

The ID of the experiment that owns the model.

namestring

The name of the model.

creation_timestamp_msint64

The timestamp when the model was created in milliseconds since the UNIX epoch.

last_updated_timestamp_msint64

The timestamp when the model was last updated in milliseconds since the UNIX epoch.

artifact_uristring

The URI of the directory where model artifacts are stored.

statusstring

The status of whether or not the model is ready for use.

Values: LOGGED_MODEL_STATUS_UNSPECIFIED, LOGGED_MODEL_PENDING, LOGGED_MODEL_READY, LOGGED_MODEL_UPLOAD_FAILED

creator_idint64

The ID of the user or principal that created the model.

model_typestring

The type of model, such as "Agent", "Classifier", "LLM".

source_run_idstring

The ID of the run that created the model.

status_messagestring

Details on the current model status.

tagsarray of object

Mutable string key-value pairs set on the model.

Show child attributesHide child attributes

keystring

The tag key.

valuestring

The tag value.

dataobject

The params and metrics attached to the logged model.

Show child attributesHide child attributes

paramsarray of object

Immutable string key-value pairs of the model.

Show child attributesHide child attributes

keystring

The key identifying this param.

valuestring

The value of this param.

metricsarray of object

Performance metrics linked to the model.

Show child attributesHide child attributes

keystring

The key identifying the metric.

valuedouble

The value of the metric.

timestampint64

The timestamp at which the metric was recorded.

stepint64

The step at which the metric was logged.

Default: 0

dataset_namestring

The name of the dataset associated with the metric. E.g. “my.uc.table@2” “nyc-taxi-dataset”, “fantastic-elk-3”

dataset_digeststring

The dataset digest of the dataset associated with the metric, e.g. an md5 hash of the dataset that uniquely identifies it within datasets of the same name.

model_idstring

The ID of the logged model or registered model version associated with the metric, if applicable.

run_idstring

The ID of the run containing the metric.

View as Markdown

The LoggedModel object

{
  "info": {
    "model_id": "string",
    "experiment_id": "string",
    "name": "string",
    "creation_timestamp_ms": 0,
    "last_updated_timestamp_ms": 0,
    "artifact_uri": "string",
    "status": "string",
    "creator_id": 0,
    "model_type": "string",
    "source_run_id": "string",
    "status_message": "string",
    "tags": [
      {}
    ]
  },
  "data": {
    "params": [
      {}
    ],
    "metrics": [
      {}
    ]
  }
}

Get GA

GET /api/2.0/mlflow/logged-models/{model_id}

Get a logged model.

API scopes: mlflow

Parameters

model_idstringpath

The ID of the logged model to retrieve.

Response

modelobject

The retrieved logged model.

Show child attributesHide child attributes

infoobject

The logged model attributes such as model ID, status, tags, etc.

Show child attributesHide child attributes

model_idstring

The unique identifier for the logged model.

experiment_idstring

The ID of the experiment that owns the model.

namestring

The name of the model.

creation_timestamp_msint64

The timestamp when the model was created in milliseconds since the UNIX epoch.

last_updated_timestamp_msint64

The timestamp when the model was last updated in milliseconds since the UNIX epoch.

artifact_uristring

The URI of the directory where model artifacts are stored.

statusstring

The status of whether or not the model is ready for use.

Values: LOGGED_MODEL_STATUS_UNSPECIFIED, LOGGED_MODEL_PENDING, LOGGED_MODEL_READY, LOGGED_MODEL_UPLOAD_FAILED

creator_idint64

The ID of the user or principal that created the model.

model_typestring

The type of model, such as "Agent", "Classifier", "LLM".

source_run_idstring

The ID of the run that created the model.

status_messagestring

Details on the current model status.

tagsarray of object

Mutable string key-value pairs set on the model.

Show child attributesHide child attributes

keystring

The tag key.

valuestring

The tag value.

dataobject

The params and metrics attached to the logged model.

Show child attributesHide child attributes

paramsarray of object

Immutable string key-value pairs of the model.

Show child attributesHide child attributes

keystring

The key identifying this param.

valuestring

The value of this param.

metricsarray of object

Performance metrics linked to the model.

Show child attributesHide child attributes

keystring

The key identifying the metric.

valuedouble

The value of the metric.

timestampint64

The timestamp at which the metric was recorded.

stepint64

The step at which the metric was logged.

Default: 0

dataset_namestring

The name of the dataset associated with the metric. E.g. “my.uc.table@2” “nyc-taxi-dataset”, “fantastic-elk-3”

dataset_digeststring

The dataset digest of the dataset associated with the metric, e.g. an md5 hash of the dataset that uniquely identifies it within datasets of the same name.

model_idstring

The ID of the logged model or registered model version associated with the metric, if applicable.

run_idstring

The ID of the run containing the metric.

View as Markdown

Response

{
  "model": {
    "info": {},
    "data": {}
  }
}

Create GA

POST /api/2.0/mlflow/logged-models

Create a logged model.

API scopes: mlflow

Request body

experiment_idstring

The ID of the experiment that owns the model.

namestring

The name of the model (optional). If not specified one will be generated.

model_typestring

The type of the model, such as "Agent", "Classifier", "LLM".

source_run_idstring

The ID of the run that created the model.

paramsarray of object

Parameters attached to the model.

Show child attributesHide child attributes

keystring

The key identifying this param.

valuestring

The value of this param.

tagsarray of object

Tags attached to the model.

Show child attributesHide child attributes

keystring

The tag key.

valuestring

The tag value.

Response

modelobject

The newly created logged model.

Show child attributesHide child attributes

infoobject

The logged model attributes such as model ID, status, tags, etc.

Show child attributesHide child attributes

model_idstring

The unique identifier for the logged model.

experiment_idstring

The ID of the experiment that owns the model.

namestring

The name of the model.

creation_timestamp_msint64

The timestamp when the model was created in milliseconds since the UNIX epoch.

last_updated_timestamp_msint64

The timestamp when the model was last updated in milliseconds since the UNIX epoch.

artifact_uristring

The URI of the directory where model artifacts are stored.

statusstring

The status of whether or not the model is ready for use.

Values: LOGGED_MODEL_STATUS_UNSPECIFIED, LOGGED_MODEL_PENDING, LOGGED_MODEL_READY, LOGGED_MODEL_UPLOAD_FAILED

creator_idint64

The ID of the user or principal that created the model.

model_typestring

The type of model, such as "Agent", "Classifier", "LLM".

source_run_idstring

The ID of the run that created the model.

status_messagestring

Details on the current model status.

tagsarray of object

Mutable string key-value pairs set on the model.

Show child attributesHide child attributes

keystring

The tag key.

valuestring

The tag value.

dataobject

The params and metrics attached to the logged model.

Show child attributesHide child attributes

paramsarray of object

Immutable string key-value pairs of the model.

Show child attributesHide child attributes

keystring

The key identifying this param.

valuestring

The value of this param.

metricsarray of object

Performance metrics linked to the model.

Show child attributesHide child attributes

keystring

The key identifying the metric.

valuedouble

The value of the metric.

timestampint64

The timestamp at which the metric was recorded.

stepint64

The step at which the metric was logged.

Default: 0

dataset_namestring

The name of the dataset associated with the metric. E.g. “my.uc.table@2” “nyc-taxi-dataset”, “fantastic-elk-3”

dataset_digeststring

The dataset digest of the dataset associated with the metric, e.g. an md5 hash of the dataset that uniquely identifies it within datasets of the same name.

model_idstring

The ID of the logged model or registered model version associated with the metric, if applicable.

run_idstring

The ID of the run containing the metric.

View as Markdown

Request

{
  "experiment_id": "string",
  "name": "string",
  "model_type": "string",
  "source_run_id": "string",
  "params": [
    {
      "key": "string",
      "value": "string"
    }
  ],
  "tags": [
    {
      "key": "string",
      "value": "string"
    }
  ]
}

Response

{
  "model": {
    "info": {},
    "data": {}
  }
}

Delete GA

DELETE /api/2.0/mlflow/logged-models/{model_id}

Delete a logged model.

API scopes: mlflow

Parameters

model_idstringpath

The ID of the logged model to delete.

View as Markdown

Response

{}

Finalize Logged Model GA

PATCH /api/2.0/mlflow/logged-models/{model_id}

Finalize a logged model.

API scopes: mlflow

Parameters

model_idstringpath

The ID of the logged model to finalize.

Request body

statusstring

Whether or not the model is ready for use. "LOGGED_MODEL_UPLOAD_FAILED" indicates that something went wrong when logging the model weights / agent code.

Values: LOGGED_MODEL_STATUS_UNSPECIFIED, LOGGED_MODEL_PENDING, LOGGED_MODEL_READY, LOGGED_MODEL_UPLOAD_FAILED

Response

modelobject

The updated logged model.

Show child attributesHide child attributes

infoobject

The logged model attributes such as model ID, status, tags, etc.

Show child attributesHide child attributes

model_idstring

The unique identifier for the logged model.

experiment_idstring

The ID of the experiment that owns the model.

namestring

The name of the model.

creation_timestamp_msint64

The timestamp when the model was created in milliseconds since the UNIX epoch.

last_updated_timestamp_msint64

The timestamp when the model was last updated in milliseconds since the UNIX epoch.

artifact_uristring

The URI of the directory where model artifacts are stored.

statusstring

The status of whether or not the model is ready for use.

Values: LOGGED_MODEL_STATUS_UNSPECIFIED, LOGGED_MODEL_PENDING, LOGGED_MODEL_READY, LOGGED_MODEL_UPLOAD_FAILED

creator_idint64

The ID of the user or principal that created the model.

model_typestring

The type of model, such as "Agent", "Classifier", "LLM".

source_run_idstring

The ID of the run that created the model.

status_messagestring

Details on the current model status.

tagsarray of object

Mutable string key-value pairs set on the model.

Show child attributesHide child attributes

keystring

The tag key.

valuestring

The tag value.

dataobject

The params and metrics attached to the logged model.

Show child attributesHide child attributes

paramsarray of object

Immutable string key-value pairs of the model.

Show child attributesHide child attributes

keystring

The key identifying this param.

valuestring

The value of this param.

metricsarray of object

Performance metrics linked to the model.

Show child attributesHide child attributes

keystring

The key identifying the metric.

valuedouble

The value of the metric.

timestampint64

The timestamp at which the metric was recorded.

stepint64

The step at which the metric was logged.

Default: 0

dataset_namestring

The name of the dataset associated with the metric. E.g. “my.uc.table@2” “nyc-taxi-dataset”, “fantastic-elk-3”

dataset_digeststring

The dataset digest of the dataset associated with the metric, e.g. an md5 hash of the dataset that uniquely identifies it within datasets of the same name.

model_idstring

The ID of the logged model or registered model version associated with the metric, if applicable.

run_idstring

The ID of the run containing the metric.

View as Markdown

Request

{
  "status": "string"
}

Response

{
  "model": {
    "info": {},
    "data": {}
  }
}

Search Logged Models GA

POST /api/2.0/mlflow/logged-models/search

Search for Logged Models that satisfy specified search criteria.

API scopes: mlflow

Request body

experiment_idsarray of string

The IDs of the experiments in which to search for logged models.

filterstring

A filter expression over logged model info and data that allows returning a subset of logged models. The syntax is a subset of SQL that supports AND'ing together binary operations.

Example: params.alpha < 0.3 AND metrics.accuracy > 0.9.

datasetsarray of object

List of datasets on which to apply the metrics filter clauses. For example, a filter with metrics.accuracy > 0.9 and dataset info with name "test_dataset" means we will return all logged models with accuracy > 0.9 on the test_dataset. Metric values from ANY dataset matching the criteria are considered. If no datasets are specified, then metrics across all datasets are considered in the filter.

Show child attributesHide child attributes

dataset_namestring

The name of the dataset.

dataset_digeststring

The digest of the dataset.

max_resultsint32

The maximum number of Logged Models to return. The maximum limit is 50.

Default: 50

order_byarray of object

The list of columns for ordering the results, with additional fields for sorting criteria.

Show child attributesHide child attributes

field_namestring

The name of the field to order by, e.g. "metrics.accuracy".

ascendingboolean

Whether the search results order is ascending or not.

Default: true

dataset_namestring

If field_name refers to a metric, this field specifies the name of the dataset associated with the metric. Only metrics associated with the specified dataset name will be considered for ordering. This field may only be set if field_name refers to a metric.

dataset_digeststring

If field_name refers to a metric, this field specifies the digest of the dataset associated with the metric. Only metrics associated with the specified dataset name and digest will be considered for ordering. This field may only be set if dataset_name is also set.

page_tokenstring

The token indicating the page of logged models to fetch.

Response

Returns a list of LoggedModel objects.

View as Markdown

Request

{
  "experiment_ids": [
    "string"
  ],
  "filter": "string",
  "datasets": [
    {
      "dataset_name": "string",
      "dataset_digest": "string"
    }
  ],
  "max_results": 0,
  "order_by": [
    {
      "field_name": "string",
      "ascending": true,
      "dataset_name": "string",
      "dataset_digest": "string"
    }
  ],
  "page_token": "string"
}

Response

{
  "models": [
    {
      "info": {},
      "data": {}
    }
  ],
  "next_page_token": "string"
}

Set Logged Model Tags GA

PATCH /api/2.0/mlflow/logged-models/{model_id}/tags

Set tags for a logged model.

API scopes: mlflow

Parameters

model_idstringpath

The ID of the logged model to set the tags on.

Request body

tagsarray of object

The tags to set on the logged model.

Show child attributesHide child attributes

keystring

The tag key.

valuestring

The tag value.

View as Markdown

Request

{
  "tags": [
    {
      "key": "string",
      "value": "string"
    }
  ]
}

Response

{}

Delete Logged Model Tag GA

DELETE /api/2.0/mlflow/logged-models/{model_id}/tags/{tag_key}

Delete a tag on a logged model.

API scopes: mlflow

Parameters

model_idstringpath

The ID of the logged model to delete the tag from.

tag_keystringpath

The tag key.

View as Markdown

Response

{}

Log Logged Model Params GA

POST /api/2.0/mlflow/logged-models/{model_id}/params

Logs params for a logged model. A param is a key-value pair (string key, string value). Examples include hyperparameters used for ML model training. A param can be logged only once for a logged model, and attempting to overwrite an existing param with a different value will result in an error

API scopes: mlflow

Parameters

model_idstringpath

The ID of the logged model to log params for.

Request body

paramsarray of object

Parameters to attach to the model.

Show child attributesHide child attributes

keystring

The key identifying this param.

valuestring

The value of this param.

View as Markdown

Request

{
  "params": [
    {
      "key": "string",
      "value": "string"
    }
  ]
}

Response

{}