ankile robust-rearrangement issues

ankile / robust-rearrangement

From Imitation to Refinement -- Residual RL for Precise Visual Assembly

https://residual-assembly.github.io/

MIT License

55 stars 3 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Seeking advice on some questions about the policy

#35 le-wei opened 1 week ago
2
Add a Mujoco environment that features a bimanual Franka setup and a simple plate insertion task

#34 ankile closed 1 week ago
0
Are there any trained models that support input of images and robot states?

#33 letsgofeng opened 1 month ago
1
WIP: Compare code from ol scaling runs to now

#32 ankile closed 1 week ago
0
Hyperparameters for reproducing results in paper

#31 edwardjjj opened 2 months ago
1
Implement chunked residual ablation

#30 ankile opened 2 months ago
0
h100 issue with isaacgym

#29 edwardjjj closed 2 months ago
0
Merge real

#28 anthonysimeonov opened 3 months ago
0
Implement state based dagger

#27 ankile closed 3 months ago
0
How to reproduce the experiments in the simulation environment

#26 yangsizhe closed 2 months ago
4
Guide to Project Workflow

#25 Henry-Ellis closed 1 month ago
3
Add an integrated residual actor with MLP as base model and enable training in same residual PPO scripts

#24 ankile closed 3 months ago
0
Update the code for the MLP RL baseline to be more aligned with latest

#23 ankile opened 3 months ago
0
Auxiliary pose prediction

#22 ankile closed 1 week ago
0
Code for VAS/IDQL baseline runs

#21 ankile closed 2 months ago
1
Add multi-gpu training capability for BC

#20 ankile closed 3 months ago
0
Relative actions

#19 anthonysimeonov closed 3 months ago
0
Real to sim

#18 anthonysimeonov closed 4 months ago
1
Fix up in handling of different parameters for different optimizers a…

#17 ankile closed 4 months ago
0
Compare file changes between when BC surely worked and current state for debugging

#16 ankile closed 4 months ago
1
Isaac sim

#15 anthonysimeonov closed 5 months ago
0
vectorized reward comp for other tasks

#14 anthonysimeonov closed 5 months ago
0
Fix: 11 upstream training changes to avoid hacky solution bc

#13 ankile closed 5 months ago
0
Vectorize reward check++

#12 ankile closed 6 months ago
0
Upstream training changes to avoid hacky solution BC + PPO for evaluation script

#11 anthonysimeonov closed 5 months ago
1
BC residual diffusion pretraining

#10 ankile closed 6 months ago
0
Actor obs queue internal and residual PPO improvements

#9 ankile closed 6 months ago
0
Align sim with real through robot adjustments and domain randomization

#8 ankile closed 6 months ago
0
Add interoperability with diffusion-adapt code

#7 ankile closed 6 months ago
0
Fix bug with number of sim steps we do that made sim time too slow

#6 ankile closed 6 months ago
0
Bring the real world eval code into the main

#5 ankile closed 6 months ago
0
Implement the residual PPO formulation

#4 ankile closed 6 months ago
0
Better action chunk prediction and cleaner actor classes

#3 ankile closed 6 months ago
0
State based rl

#2 ankile closed 6 months ago
0
Implement state-based bc training for MLP and diffusion

#1 ankile closed 7 months ago
0