maxnleo

Question about loss‐scaling factors (e.g., 0.01, 1e2) and penalty parameters (mu_PDE, mu_V) in your paper

1

Hello, Dear @lululxvi @rpestourie @WenjieYao thanks for making the hpinn code available. I found the paper and code very helpful. However, in stoke flow example, I am facing some difficulties...

General Guidelines for Scaling PDE Residuals and Weighting Hard Constraints in Augmented Lagrangian PINNs

1

Hello @lululxvi, I have reviewed the FAQ discussions related to *"I failed to train the network or get the right solution, e.g., large training loss, unbalanced losses."* but I'm still...