search

gsartoretti / PRIMAL

PRIMAL: Pathfinding via Reinforcement and Imitation Multi-Agent Learning -- Distributed RL/IL code for Multi-Agent Path Finding (MAPF)
MIT License
294 stars 78 forks source link

Inquire for Programs #2

Open ghost opened 5 years ago

ghost commented 5 years ago

Dear Dr. Sartoretti:

I am a research assistant from The Hong Kong Polytechnic University, and I'm really interested in your work: multi-agent path finding. We are using agents for some constructions in building design.

When I run your programs, the agents are initialized, while the GPU information show no training progress. It seems that the program just stops at the initializing stage. Would you know how to solve this problem?

I am looking forward to your response.

Thanks & regards, Pengkun Liu

image

微信图片_20190611172427

wanggsz commented 1 year ago

Hi I have the same question. After the training process starting worker 24, it reported terminate called after throwing an instance of 'std::bad_alloc' what(): std::bad_alloc Aborted (core dumped) I wonder if you have solved the problem and what's the solution. Thank you so much for reading this and waiting for your reply!

Jing-Loog commented 1 year ago

@wanggsz Your computer hardware limits the running of the code, reduce the number of workers and metaAgent, and reduce the size of the agent field of view and network size