get a small loss but the results are completely not correct

[weiroy]

Hi! @lululxvi Thanks very much for your amazing work!
But I encountered a problem. Specifically, the loss is very small, but the results are completely not correct. I am wondering why this happend.

The loss is about 1e-5, in figure 2, the black line is the reference solution and the red line the prediction results of deepxde.
And the governing equations, boundary conditions are as follows:

Looking forward to your reply.

[BraveDrXuTF]

Are you using tf1 backend? I got similar issue. See here

[yuluoxiansen]

Have you solved it? I have the same problem.

[lululxvi]

Is your problem well defined?

[weiroy]

It's my mistake. The right ode system is as follows:

But I encountered another problem. I have non-dimensionalized the ode system. And the coefficient 0.001 can not be converted to the order of 1 by applying some scaling tricks to the ode system. Some times, the loss is large and the results of pinn are not equal to the reference solution. I guess this is because of the coefficient 0.001, but I don't know how to solve this problem.