Migrating from skrl¶
skrl’s strength is algorithm breadth. UniLab focuses on a curated set (PPO, SAC, TD3 with several optimized variants) but adds a real-hardware deployment path.
Mapping skrl concepts to UniLab¶
skrl |
UniLab |
|---|---|
|
|
|
|
|
|
|
|
|
|
What to expect¶
No algorithm parity for niche algos (CQL, IQL, etc.) — UniLab intentionally focuses on a few highly optimized actor-critic variants.
Different runner lifecycle. skrl’s monolithic trainer becomes a collector + learner pair connected by shared memory. See Runner Lifecycle.
Different env interface. skrl tolerates many env styles. UniLab insists on
NpEnv+ dict obs.
Migration checklist¶
Decide which UniLab algorithm best matches your skrl agent.
Port the env into
NpEnvform.Convert hyperparameter YAML into Hydra groups under
conf/<algo>/<task>/.Validate reward parity.