yunshengtian / Assemble-Them-All

[SIGGRAPH Asia 2022] Assemble Them All: Physics-Based Planning for Generalizable Assembly by Disassembly
http://assembly.csail.mit.edu
MIT License
139 stars 15 forks source link

void redmax::Simmulation::forward(int, bool, bool):Assertion (q - _q_his[_q_his.size() - 1]).norm() < 1e-7 && (qdot - _qdot_his[_qdot_his.size() - 1]).norm() < 1e-7; failed 已放弃(核心已转储) #10

Closed Zlzxzw1234 closed 7 months ago

yunshengtian commented 8 months ago

Hello @Zlzxzw1234 let me know if you have specific questions. Thanks

taniguchi-t-nmri commented 8 months ago

Sorry for interrupting... I also had the same fail, so let me describe it here.

Execution environment:

When the fail occurs:

However, the 00004th in the "joint_assembly" directory is successfully computed. """ python examples/run_joint_plan.py --planner bfs --dir joint_assembly --id 00004 Status: Success, planning time: 12.665902853012085 """

(I am very impressed with your work.The follow-up "ASAP" is also excellent.)

yunshengtian commented 8 months ago

Thanks @taniguchi-t-nmri for providing more details of the problem!

I tried it on my laptop, and I couldn't reproduce this error. I am using MacBook M1. They were also significantly faster than yours somehow (test_joint_sim.py on 00016 takes 2.83s and run_joint_plan.py on 00016 takes 0.28s).

Let me try later to see if I can reproduce on my linux machine. But anyway, this is a very weird issue that shouldn't happen. I will keep you posted.

yunshengtian commented 7 months ago

I tried on my linux machine and still couldn't replicate this issue. I would suggest a simple fix - just comment out the assertion line and rebuild the simulation using python setup.py install and see if you can still get something reasonable. Otherwise, we may need to delve deep into the code and check why the assertion has failed.