NOTE: Using experimental fast data loading logic. To disable, pass "--load_fast=false" and report issues on GitHub

[nga77]

I keep getting this message when running the lcd example in Modulus.
The tensorboard link does not work and I have tried using the command in the github page.

How else do I go about this so I can visualize the results?

[ngeneva]

Tensorboard is the primary method for visualizing Modulus results. I would suggest looking online for tensorboard solutions to help solve your problem. Alternatively, one can always modify Modulus to write out losses and plots to file:

Plotting of training and validation losses

Hi @id21resch11019 Presently, no. Tensorboard is our built in logging backend in Modulus. However customizing Modulus to log any way you want is straight forward. In the trainer.py, you can add a logging method of your choice based on the summary_freq. The losses for a given iteration are stored in the losses dictionary.