LogisticRegression model not achived

[KaiyiLiu1234]

What happened?

For Manual Metric Collection and Training with Entrypoint, all models produced appears to have an extracted folder and a zipped folder with the exception of LogisticRegression.

What did you expect to happen?

LogisticRegression should also have an archive.

How can we reproduce it (as minimally and precisely as possible)?

Run Manual Metric Collection and Training with Entrypoint and view rapl-sysfs.

Anything else we need to know?

No response

Kepler image tag

latest

Deployment

• Model server
• Estimator
• Online trainer
• Offline trainer
• Profiler

Kepler model server image tag if deployed

Kepler estimator image tag if deployed

Kepler online trainer image tag if deployed

Kepler offline trainer image tag if deployed

Kepler profiler image tag if deployed

Kubernetes version

$ kubectl version
# paste output here

Install tools

Kepler deployment config

For on kubernetes:

$ KEPLER_NAMESPACE=kepler

# provide kepler configmap
$ kubectl get configmap kepler-cfm -n ${KEPLER_NAMESPACE} 
# paste output here

# provide kepler model server configmap if Kepler Model Server is deployed 
$ kubectl get configmap kepler-model-server-cfm -n ${KEPLER_NAMESPACE} 
# paste output here

# provide kepler deployment description
$ kubectl describe deployment kepler-exporter -n ${KEPLER_NAMESPACE} 

For standalone:

put your Kepler command argument here

[sunya-ch]

@KaiyiLiu1234
Have you checked the error threshold?
We have a condition to select only trained model that pass the error threshold here:

kepler-model-server/src/kepler_model/train/exporter/validator.py

Lines 101 to 104 in a0ad11f

	if row["mae"] > hard_threshold or row["mape"] > hard_threshold:
	continue
	if row["mape"] <= mape_threshold or row["mae"] <= mae_threshold:
	model_name = row["model_name"]

[KaiyiLiu1234]

I see. I think LogisticRegression error is always too high.