def run_train(
forecaster: Estimator,
train_dataset: Dataset,
hyperparameters: dict,
validation_dataset: Optional[Dataset],
) -> Predictor:
num_workers = (
int(hyperparameters["num_workers"])
if "num_workers" in hyperparameters.keys()
else None
)
shuffle_buffer_length = (
int(hyperparameters["shuffle_buffer_length"])
if "shuffle_buffer_length" in hyperparameters.keys()
else None
)
num_prefetch = (
int(hyperparameters["num_prefetch"])
if "num_prefetch" in hyperparameters.keys()
else None
)
if isinstance(forecaster, GluonEstimator):
return forecaster.train(
training_data=train_dataset,
validation_data=validation_dataset,
num_workers=num_workers,
num_prefetch=num_prefetch,
shuffle_buffer_length=shuffle_buffer_length,
)
else:
return forecaster.train(
training_data=train_dataset, validation_data=validation_dataset,
)
def test_item_id_info(dataset: Dataset, estimator: Estimator):
predictor = estimator.train(dataset)
forecasts = predictor.predict(dataset)
for data_entry, forecast in zip(dataset, forecasts):
assert (not "item_id"
in data_entry) or data_entry["item_id"] == forecast.item_id
assert (not "info"
in data_entry) or data_entry["info"] == forecast.info
def run_train(
forecaster: Estimator,
train_dataset: Dataset,
validation_dataset: Optional[Dataset],
) -> Predictor:
log.metric("train_dataset_stats", train_dataset.calc_stats())
return forecaster.train(train_dataset, validation_dataset)
def run_train(
forecaster: Estimator,
train_dataset: Dataset,
validation_dataset: Optional[Dataset],
) -> Predictor:
return forecaster.train(
training_data=train_dataset, validation_data=validation_dataset
)