diff --git a/docs/_layouts/home.html b/docs/_layouts/home.html index 8aa8cc4..5911eca 100644 --- a/docs/_layouts/home.html +++ b/docs/_layouts/home.html @@ -34,11 +34,34 @@ + + @@ -367,7 +390,7 @@
By randomly initializing the object layouts and running the object-aware retargeting pipeline each time, we can efficiently generate a large volume of successful rollout data using OKAMI without the need for human-teleoperation. The rollout data can then be used to train closed-loop visuomotor policies through behavioral cloning. We test on two tasks, bagging and sprinkle-salt. The success rates of the visuomotor policies achieve 83.3% and 75% respectively. Here we provide the rollouts of visuomotor policies.
@@ -511,7 +534,7 @@OKAMI's policies may fail to grasp objects due to inaccuracies in the robot controllers, the human reconstruction model or the vision models, or fail to complete tasks because of unwanted collisions, undesired upper body rotations, or inaccuracy in solving inverse kinematics. Here we provide typical failure examples.
-