ClearML gets double variables, why?

Any idea why ClearML outputs this variable eps_0 as double? (with eps_1 as well)?

I’m using TensorBoardCallback() as a parameter to the argument cbs.

I think it would be much better to ask on their slack channel or make issue on github with more details, as this forum has little relation to the ClearML project

Thanks! You’re right. I’ll ask them.
I’m asking here because I was integrating ClearML with fast.ai, and maybe these scalars are exclusive to fast.ai.

Hello,
ClearML outputs eps_0 as a double because it’s designed to handle floating-point precision for epsilon values, which is common in machine learning frameworks. The TensorBoardCallback logs these values for visualization and analysis purposes.FLVSlogin Net

Best Regards,
Joan Martin