Caltech-AMBER / ambersim

In-house tools built on GPU-accelerated simulation
MIT License
7 stars 2 forks source link

Save and load policies from a file #29

Closed vincekurtz closed 10 months ago

vincekurtz commented 10 months ago

The MJX tutorial has a partial example, but it's incomplete, as it relies on a make_inference_fn that was created during training and never saved. We should have some utilities to make saving and loading easy, particularly for non-default policy architectures.

alberthli commented 10 months ago

Whoever picks this up should start from here: https://flax.readthedocs.io/en/latest/api_reference/flax.serialization.html