search

lucys0 / awe

Waypoint-Based Imitation Learning for Robotic Manipulation
https://lucys0.github.io/awe/
73 stars 7 forks source link

i get bad performance #4

Open deavn2236 opened 4 months ago

deavn2236 commented 4 months ago

Hello, after I use waypoint to train in the simulation environment, I find that the success rate of using waypoint is 90%, while the success rate of not using waypoint is 98%, which decreases the success rate. May I ask why this is? Besides, the imitation learning performance is also related to the number of exploration states. Doesn't it affect his state space when using waypoint

Bailey-24 commented 1 month ago

I can't find the task, which make coffee in the real world, do you find it?

Because I think this work is very similar to Figure01 robot making coffee.