What's the expected result from examples/Burgers.py?

[piyueh]

Hi, I'm trying the code in examples folder to make sure my environment setup is correct (e.g., dependencies, versions, etc.). I tried Burgers.py. It ran without error messages (except some deprecation warnings from TensorFlow). But how to know whether the result is what I should expect?

Or if there's something like unit tests for users to know whether the installation works correctly?

This is the figure I got from Burgers.py:

Thanks!!

[lululxvi]

Your result is correct. If the code can successfully run without error, then the setup should be correct.

For Burgers, the input variable is two dimensional (x, t), and the output is y, so the function surface is 3D. But, by default, the code projects the 3D plot into a 2D figure by ignore t. That is why the plot looks messy. The result is saved to test.dat with the first column x, the second column t, the third column y_true, and the fourth column y_pred. You can plot the result in 3D to compare the exact solution and prediction.

You can also try simples cases, such as Poisson in 1D.