Transfer learning

Physics Simulation Modulus Physics-ML Model Framework Technical Support (Modulus Only)
1

Hello,

I am trying to use transfer learning and just wanted to confirm that I am doing it correctly. I previously created a model and stored the weights in a folder. I then updated my new model and added to the configuration file

initialization_network_dir : “outputs/previous_model”

Are there anymore steps that are required? I asked because this model is taking the exact amount of time and epochs compared to previous model even given it the exact same results.

Thank you for your help.

2

Solved. The issue was that since the model is saved in the relative path of “outputs” it cannot see the other models in outputs. Using the to_absolute_path function helped or I could of put the absolute path in the network directory.

1 Like
3

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.