When I train models in the serverless environment V4 (Premium Plan), the system occasionally returns the error message listed below, especially after running the model training code multiple times. We have tried creating new serverless sessions, which sometimes helps but not consistently. We also followed the suggestion in the error message and deleted previously trained models, but that did not resolve the issue. This behavior also occurs in the Free Edition.

"[CONNECT_ML. MODEL_SIZE_OVERFLOW_EXCEPTION] Generic Spark Connect ML error. The fitted or loaded model size is about 572745576 bytes. Please fit or load a model smaller than 268435456 bytes. SQLSTATE: XX000"

Based on some preliminary research, I suspect this may be related to incomplete memory cleanup or garbage collection. We would greatly appreciate it if you could check whether there is a known solution or connect us with the appropriate person to help investigate further.