prediction module¶
-
class
prediction.
Prediction
(client, uuid)¶ Bases:
object
-
property
current_results
¶ ids of current results
-
delete
()¶ Delete this prediction job
-
download
(save_dir=None)¶ Download the prediction results as tar.gz archive.
- Args:
save_dir: directory to save the results.
Returns:
-
get_results
(sample_ids)¶ - Args:
sample_ids (list): id(s) of the sample(s) to retrieve
- Returns:
AttrDict of numpy array: sample (including ground truth if present), and prediction
-
property
info
¶ Prediction job info
-
restart
()¶ Restart this prediction job
-
stop
()¶ Stop this prediction job
-
static
submit
(client, trained_model_id, dataset_id)¶ Submit a new batch prediction job
- Args:
client: API client trained_model_id: trained model id to use dataset_id: dataset id to use
- Returns:
Prediction
-
property