Client

Client represents a higher level interface to training API.

class abeja.train.Client(organization_id: Optional[str] = None, job_definition_name: Optional[str] = None, training_job_id: Optional[str] = None, credential: Optional[Dict[str, str]] = None, timeout: Optional[int] = None, max_retry_count: Optional[int] = None)

A High-Level client for Training API

from abeja.train import Client

client = Client(organization_id='1234567890123')

Params:

• organization_id (str): The organization ID. Takes from os.environ[‘ABEJA_ORGANIZATION_ID’] if omitted.

• job_definition_name (str): The job definition name. Takes from os.environ[‘TRAINING_JOB_DEFINITION_NAME’] if omitted.

• training_job_id (str): The training job ID. Takes from os.environ[‘TRAINING_JOB_ID’] if omitted.

download_training_result(training_job_id: Optional[str] = None, path: Optional[str] = None) → None

download training artifact that includes model file or logs. Training artifact itself is a zip file, this function extracts it.

API reference: GET /organizations/<organization_id>/training/definitions/<job_definition_name>/jobs/<training_job_id>/result

Request Syntax:

training_job_id = '1600000000000'
client.download_training_result(training_job_id)

Params:

• path (str): target path where result are extracted [optional]

Raises:

• abeja.exceptions.BadRequest

• abeja.exceptions.Unauthorized

• abeja.exceptions.Forbidden

• abeja.exceptions.NotFound

• abeja.exceptions.MethodNotAllowed

• abeja.exceptions.InternalServerError

• json.JSONDecodeError

• PermissionError

• ValueError

• IOError

update_statistics(statistics: abeja.training.statistics.Statistics) → None

Notify a job statistics for ABEJA Platform.

API reference: POST /organizations/<organization_id>/training/definitions/<job_definition_name>/jobs/<training_job_id>/statistics

Request Syntax:

from abeja.train import Client
from abeja.train.statistics import Statistics as ABEJAStatistics

client = Client()

statistics = ABEJAStatistics(num_epochs=10, epoch=1)
statistics.add_stage(name=ABEJAStatistics.STAGE_TRAIN, accuracy=90.0, loss=0.10)
statistics.add_stage(name=ABEJAStatistics.STAGE_VALIDATION, accuracy=75.0, loss=0.07)

client.update_statistics(statistics)

Params:

• statistics (abeja.train.statistics.Statistics): job statistics to nofity

Returns:

None

Statistics

class abeja.train.statistics.Statistics(num_epochs: Optional[int] = None, epoch: Optional[int] = None, progress_percentage: Optional[float] = None, **kwargs)

STAGE_TRAIN = 'train'

STAGE_VALIDATION = 'validation'

add_stage(name: str, accuracy: Optional[float] = None, loss: Optional[float] = None, **kwargs) → None

add stage information

Params:

• name (str): name of stage. It have prepared STAGE_TRAIN and STAGE_VALIDATION as constants, but you can set arbitrary character strings.

• accuracy (float): accuracy rate that value needs between 0 and 1.

• loss (float): loss rate that value needs between 0 and 1.

Returns:

None

Raises:

• ValueError

classmethod from_response(response: Optional[Dict[str, Any]]) → Optional[abeja.training.statistics.Statistics]

get_statistics() → dict

get stage information

Return type:

dict

Returns:

Response Syntax:

{
    'num_epochs': 10,
    'epoch': 1,
    'progress_percentage': 90,
}