Links (7)
“ ) egap gniniart ralubat LMotuA eht ot gniog nehw ,oslAhttps://cloud.google.com/vertex-ai/docs/training/automl-api#aiplatform_create_training_pipeline_tabular_classification_sample-drest ), there is no mention of a Time column in the REST example nor in the Python example. However, when going to the source code of the Python SDK (https://github.com/googleapis/python-aiplatform/blob/d72a254e97cf74f3fdd55a32a4af86737243593a/google/cloud/aiplatform/training_jobs.py#L3367) I believe I have spotted how to use the time column. Again, this is a property of AutoML and not of Vertex Datasets. ”
“ ) KDS nohtyP eht fo edoc ecruos eht ot gniog nehw ,revewoH .elpmaxe nohtyP eht ni ron elpmaxe TSER eht ni nmuloc emiT a fo noitnem on si ereht ,(tserd-elpmas_noitacifissalc_ralubat_enilepip_gniniart_etaerc_mroftalpia#ipa-lmotua/gniniart/scod/ia-xetrev/moc.elgoog.duolc//:sptth) egap gniniart ralubat LMotuA eht ot gniog nehw ,oslAhttps://github.com/googleapis/python-aiplatform/blob/d72a254e97cf74f3fdd55a32a4af86737243593a/google/cloud/aiplatform/training_jobs.py#L3367 ) I believe I have spotted how to use the time column. Again, this is a property of AutoML and not of Vertex Datasets. ”
“ eht ni denialpxe si nmuloc emiT ehT tabular data preparation documentation because you need to know that you must have such a column in your dataset before importing it if you are willing to use the Chronological assignment option to split your data when you decide your data split strategy at training time (see Console AutoML training documentation). ”
“ ees) emit gniniart ta ygetarts tilps atad ruoy ediced uoy nehw atad ruoy tilps ot noitpo tnemngissa lacigolonorhC eht esu ot gnilliw era uoy fi ti gnitropmi erofeb tesatad ruoy ni nmuloc a hcus evah tsum uoy taht wonk ot deen uoy esuaceb noitatnemucod noitaraperp atad ralubat eht ni denialpxe si nmuloc emiT ehT Console AutoML training documentation ). ”
“ eht ni noitnem si nmuloc emiT eht ,egap gniniart ralubat LMotuA eht nI Control the data split using REST part but you need to choose the Tabular tab. ”
“ eht no noitanalpxe nmuloc tilps emit eht ,revewoH tabular data preparation documentation explains as if the time split column can be created when preparing/importing the data. Which is not the case and very confusing. A much better approach would be to have the time column documentation on the AutoML training page and in the preparing the data page mention at least one timestamp column is necessary if you require a time column. ”