Unnormalize actions with model#sample_actions in OctoInference

simpler-env / SimplerEnv

Evaluating and reproducing real-world robot manipulation policies (e.g., RT-1, RT-1-X, Octo) in simulation under common setups (e.g., Google Robot, WidowX+Bridge)

https://simpler-env.github.io/

MIT License

137 stars 12 forks source link

Unnormalize actions with model#sample_actions in OctoInference #5

Closed omeryagmurlu closed 2 weeks ago

omeryagmurlu commented 2 weeks ago

added functionality to allow user-supplied OctoModel in OctoInference
unnormalize actions with model#sample_actions instead of hardcoding them in code