RLBench
RLBench copied to clipboard
stepjam
A large-scale benchmark and learning environment.
Currently, resetting is not consistent due to two things: - Sampled boundary is random and depends on global np seed (https://github.com/stepjam/RLBench/blob/master/rlbench/backend/spawn_boundary.py#L159) - max_attempts introduces more samples. Even if a global...
Hi @stepjam, is it possible to use RLBench on a machine without an X server?
How does PyRep implement inverse kinematics for robotic arms? Currently, no one has answered questions about inverse kinematics on GitHub. I hope a more detailed inverse kinematics tutorial can be...
Hi, I want to change the Panda into UR3 when using the env = gym.make('reach_target-state-v0', render_mode='human'). But it seems that even if I make changes into env = Environment( action_mode,...
Hi, I am training imitation learning policies to solve RLBench tasks. I am currently facing the following issue: even though the predicted trajectory is correct, the robot fails the task...
Run tasks under `rlbench/tasks` using `task_builder`, start simulation with `+`, and `d` for demo, but CoppeliaSim keeps showing simulation in progress, the Panda arm doesn't move, and the simulation won't...
Hi Stephen, thanks for your continued work on RLBench. And great talk at ICRA this year! Quick question: how to change the simulation time step for demo collection? I've tried...
Seems the link going to the gifs of the task is not working anymore: [https://drive.google.com/drive/folders/1TqbulbbCEqVBd6SBHatphFlUK2JQLkYu?usp=sharing](https://drive.google.com/drive/folders/1TqbulbbCEqVBd6SBHatphFlUK2JQLkYu?usp=sharing)
While using RLBench to collect data with custom camera views, I encountered a persistent black rectangular artifact in the rendered images. The black shape appears to float in the scene...
