How to stop training

[happy-machine]

Hi,

Firstly, thank you!

I've followed your docs, I'm using an AWS Ubuntu 20 AMI in headless mode.

I'm trying to run the fox example using the exact command given, and when running I assumed that it would run for a certain number of epochs or until the loss hit a threshold.Instead it carried on running all night and is still running.

I've looked at the Python options and read through the docs multiple times .. where it suggests that you somehow need to stop training to then get the predictions / output

Can you explain how this is done?

[happy-machine]

For anyone else struggling, use the Python wrapper Python scripts/run.py

Running it will give you a list of available options (would be useful for these to be in the docs)

Use --n_steps 15000 (for example) to exit training after steps in headless mode, use --save_mesh to save mesh and --save_snapshot to save the model