CODEWITHSUNDEEP
pythonmlflow
bioinformatics_student
bioinformatics_student

Reputation: 488

fitting and predicting model with mlflow

I'm very new to understanding the use of MLFlow but need assistance, I'm trying to understand on how to try and fit and predict my model once again. I'm able to call my model by:

PLS_model = mlflow.pyfunc.load_model("runs:/FFFFF!@#!@#@!#!/logged_model", suppress_warnings = True)

and get:

mlflow.pyfunc.loaded_model:
  artifact_path: logged_model
  flavor: mlflow.sklearn
  run_id: FFFFF!@#!@#@!#!

But when I try to call any methods as:

1).fit or .predict. I get the following error

AttributeError: 'PyFuncModel' object has no attribute 'fit'

AttributeError: 'PyFuncModel' object has no attribute 'predict'

Here I encountered on how to actually call these functions but not sure if I'm doing this correctly. In summary, how can I predict, fit to my new data.

Thanks

Upvotes: 1

Views: 5381

Answers (1)

Sergey Pryvala
Sergey Pryvala

Reputation: 525

Try to use sklearn model loader

import mlflow.sklearn
model = mlflow.sklearn.load_model(..)
model.predict(...)

Upvotes: 4

Related Questions