Sagemaker Autopilot Endpoint OutOfMemory

0

Good morning, I am encountering an unusual issue with Sagemaker Autopilot. After launching the training on my dataset and completing the entire process without any problems, the best model is deployed, and the endpoint appears to be functioning correctly. However, the issue arises when I attempt to call the endpoint repeatedly. I observe that the memory usage rapidly increases until it reaches 97%. At that point, it either crashes with the error message "worker died," and the CloudWatch logs indicate an "OutOfMemory" error in Java, or it remains operational but becomes congested. Upon reviewing CloudWatch, there doesn't appear to be anything abnormal except for the recurring message: "The column 'column_name' does not exist in your dataset. Please specify a different column name for the 'Input column' that actually exists in your dataset and try again." This message repeats even before I start invoking the endpoint. I'm uncertain if this is indeed the cause of the problem.

preguntada hace un año99 visualizaciones
No hay respuestas

No has iniciado sesión. Iniciar sesión para publicar una respuesta.

Una buena respuesta responde claramente a la pregunta, proporciona comentarios constructivos y fomenta el crecimiento profesional en la persona que hace la pregunta.

Pautas para responder preguntas