Default / 默认 · 8月 31, 2021




Kaggle’s GPU usage is limited to 30 hours per week for each type of processor . This means that if you use both GPUs and TPUs. There is also a limit on how much disk space you can use for your active work . The limit resets every week, but it’s unclear on which day of the week it resets.

It’s worth noting that Kaggle also provides CPU and TPU processing. Additionally, all notebooks in Kaggle timeout after 12 hours of execution , regardless of whether they are using an accelerator or not. Finally, Kaggle has recently slightly limited concurrent GPU sessions, allowing only 1 interactive GPU session and 2 commit sessions at one time , to better allocate resources.

Overall, while there are limits on GPU usage in Kaggle, it still provides a valuable resource for those looking to train machine learning models without expensive hardware.

%d 博主赞过: