What type of model did AutoML train?

[Amit Kagian]

I recently started to use autoML and I think it is great. However, I am unable to figure out what type of model did autoML eventually arrived at on my data; is it an NN? GBDT? AdaNet? 

The documentation refers me to Cloud Logging, although I am still not sure how to figure out the eventual trained model.

I searched the logs with my model_id, but I am not sure how to understand what I see. 
Is the last log entry represent the eventual model? 

In addition, for some strange reason I see only log entries for 2021-06-30 02:27:20 although the model ran for 16 hours. Does this make sense?

This is the query I used: 

resource.type="cloudml_job" resource.labels.job_id="TBL5690762123089543168" resource.labels.project_id="waze-bi" labels.log_type="automl_tables" jsonPayload."@type"="type.googleapis.com/google.cloud.automl.master.TuningTrial"

Can anybody please help?

[Chenyu Zhao]

From Yifan on the separate thread:

Yes it seems you are doing the right thing. Once you search and  find  any TuningTrial object, you can unfold it in the stackdriver UI. then you can see our hyperparameters, and there is one parameter named "Model type". Currently we are using only NN and TFBT, as they are the winners in most cases, so we currently disabled Adanet, etc.

...

to find the finally used model, you can filter the log with @type"="type.googleapis.com/google.cloud.automl.master.TablesModelStructure" instead of TuningTrial.

then that's the final ensemble model. For the tuning trials, we just output the top 200 trials we tried in the very first hyperparameter searching phase. 

--
You received this message because you are subscribed to the Google Groups "cloud-automl-tables-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cloud-automl-tables...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cloud-automl-tables-discuss/3d248a76-6f2b-4244-abfb-578e354795cen%40googlegroups.com.