You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As franka cobot price doubled, I bought another robot arm which is 6DOF.
At a glance, it appears that 7 joint position information is used only in the replay process,
but I think the information used in the deep learning model has nothing to do with 7DOF.
I'm curious about your opinions.
The text was updated successfully, but these errors were encountered:
We see lots of evidence that co-training with diverse embodiments is indeed possible, and if you collect a dataset for your robot arm and mix it with the DROID dataset, chances are you will see some transfer (ie your robot policies will work better / more robustly than when only training on the data from your robot).
However, what I would not expect to work is to use the policy trained on DROID directly on your robot, without collecting any data for your robot.
As franka cobot price doubled, I bought another robot arm which is 6DOF.
At a glance, it appears that 7 joint position information is used only in the replay process,
but I think the information used in the deep learning model has nothing to do with 7DOF.
I'm curious about your opinions.
The text was updated successfully, but these errors were encountered: