AI engineering glossary · Computing foundations
What is OOM (out of memory)?
Out of memory (also called OOM): An error state in which a program or system asks for more memory than is available and cannot get it, so the process fails or is killed. In machine learning it is common when a batch or model is too big for GPU or system memory.
Out of memory: an example
Training with batch size 512 crashes with a CUDA out-of-memory error, and reducing the batch size to 128 fixes it.
Related terms
More Computing foundations terms
Where you learn it
Out of memory is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.