Skip to content

Training Resource


This is an object representing your Training.

Each Training has a dataset as input and outputs a model.

You can retrieve it to see the state of the training.


GET /v1/trainings/

GET /v1/trainings/{TRAINING_ID}/

POST /v1/trainings/

DELETE /v1/trainings/


Top level attributes

Writable attributes

Name Type Description
source_dataset_id string The ID of the dataset that will be used as input for the training
output_model_name string The name for the model that the training will create
output_model_description string The description for the model that the training will create
output_model_changelog string The changelog for the version of the model that the training will create
config dict Specific parameters for the training (See Available training parameters)

These attributes can be sent when creating a Training object ('''POST /v1/trainings/''')

Read only attributes

Name Type Description
id string Global identifier to access the actual resource
type string The type of the input dataset (See available dataset types)
state string Current state of the training ("waiting", "processing", "completed", "failed")
error string The error that happened during processing the training
progress number The current progress of the training (0.0 - 1.0)
duration number The total or current duration of the training in seconds
time_created string Creation time of the training (ISO Time String)
time_started string Starting time of the training (ISO Time String)
time_completed string Completion time of the training (ISO Time String)
source_dataset Dataset object The input dataset for this training
output_model Model object The model that was created when the training is done, otherwise the value is null

Available training parameters (training config)

Based on the '''type''' field, config can contain different data.

Type "face"

Name Type Default Description
min_face_size integer 80 Minimum size of the smallest side of the face in pixels
sharpness_threshold integer 40 Minimum quality of the face (lower number is more blurry)

Type "landmark"

No config parameters available yet

JSON Example

The following JSON snippet is showing a completed training of a face dataset.

    "id": "a09d87ed-1c12-4ea5-827e-906634626797",
    "type": "face",
    "state": "completed",
    "progress": 1.0,
    "duration": 2.8,
    "error": "",
    "time_created": "2020-12-15 12:34:18.237000",
    "time_started": "2020-12-15 12:55:35.712000",
    "time_completed": "2020-12-15 12:56:18.601000",
    "source_dataset": {
        "id": "b97e20c8-3e68-4831-8eaf-e7a592d3a802",
        "name": "Test",
        "description": "A test dataset",
        "version": 1.0,
        "type": "face",
        "time_created": "2020-10-30 17:34:41.480000",
        "time_updated": "2020-10-30 17:34:41.480000",
        "number_of_classes": 0,
        "number_of_active_classes": 0,
        "number_of_images": 0,
        "number_of_active_images": 0,
        "preview_images": [],
        "evaluation_result": {
            "gender_statistics": {
                "male": 5
        "evaluation_state": 0,
        "issues": {
            "has_error": true,
            "has_warning": false
    "output_model_name": "My Model",
    "output_model": null,
    "config": {},
    "output_model_description": "This is a model which training was started via Postman",
    "output_model_changelog": "initial training"