Open spiglerg opened 1 week ago
RunningMeanStd is not overflow safe, and overflows when running large-scale training (e.g., on a cluster).
I'm submitting a pull request with a proposal to address the problem.
No response
I have provided a minimal and working example to reproduce the bug
hello please do not forget this step, that would allow us better to understand the problem and when it occurs.
🐛 Bug
RunningMeanStd is not overflow safe, and overflows when running large-scale training (e.g., on a cluster).
To Reproduce
I'm submitting a pull request with a proposal to address the problem.
Relevant log output / Error message
No response
System Info
No response
Checklist