| # VariBAD ODE model artifact | |
| This repository contains a deployable VariBAD ODE model artifact. | |
| - Backbone: `varibad` | |
| - ODE system: `scalar_decay` | |
| - Source training iteration: `750` | |
| - Artifact format version: `1` | |
| - Weights: `vae.safetensors`, `policy.safetensors` | |
| - Configuration: `config.yaml` | |
| - Training optimizer state and PID control-envelope calibration are not included. | |
| Load the artifact with the project's `load_model_artifact` Python interface or | |
| the `steppo-hf download` command. The loader reconstructs the VAE, policy, | |
| ODE environment, and `LearnedController` from the packaged configuration. | |