Kaustubh Tangsali
Kaustubh Tangsali
This is an expected behavior of the LBFGS optimizer in Modulus-Sym. Inside Modulus-Sym, the optimizer will set the `max_steps` to zero. If the training is started from scratch, this issue...
/blossom-ci
/blossom-ci
> @mnabian added a couple of comments which I feel will have to be addressed. Let's discuss this further @mnabian fixed the issues I pointed out and also used this...
Multi-GPU CI can be currently run on-demand (using `multi-gpu-ci`). However we are yet to integrate them in the coverage report. Considering shortage of CI hardware, I don't think we can...
Hi @azrael417 , @akshaysubr any updates on this?
Just an update on this issue, currently this can be handled via CI on the internal mirror. The team is looking into making this available on the public side.
This is implemented and functional now. Please run `/multi-gpu-ci` in comments to trigger multi GPU tests.
Hi @ankurmahesh @akshaysubr does this issue still exist?