Closed Arseni1919 closed 7 months ago
Hi! Yes, py-lacam is super slow.
This is because the performance of LaCAM depends on the configuration generator. The papers use PIBT for this, while py-lacam uses random action selection. Note that this repo just serves as a playground.
There is also a PIBT implementation in Python. By combining them, it will be possible to realize a fast version. (Still, c++ is the best for the performance) https://github.com/Kei18/pypibt
Oh.. ok! Is it possible, by any chace, that you could incorporate into LaCAM the required configuration generator (PIBT) in this repo? 🙏 So that we will not make any mistake in doing so. In our case, the algorithms, that are written in the same language are good for inter-comparison (despite the fact, that C++ is know to be much faster). Thank you a lot for a quick response!
I see. The Python implementation is possible and not a hard task. I think it only has to add about 100 lines. But I do not have the bandwidth at the moment. I cannot promise when I will do it...
Oh.. it is tottaly fine! At your earliest convenience, and thank you a lot in advance! Thank you for your response!
Okay, I will notify you when it is done.
Cheers, Keisuke
@Arseni1919 I eventually wrote a code for LaCAM-PIBT. https://github.com/Kei18/py-lacam/tree/pibt
Please note that I did not go into code-level optimization or performance optimization of the algorithm itself. It's at the proof-of-concept level.
Thank you! It really helps us a lot!:) I will try to run it shortly. Thanks 🙏
Hi! I've tried to run the
py-lacam
code on my device (Macbook Air 1M), but, unfortunatelly, the execution seems to be really slow. Below are the adjsutments that I did. The maps are from the standard benchmark. The algorithm claims to perform really fast, but, maybe, its perfomance is the best only in C++? Can it be the case? I am implementing all of the algorithms in Python, that is why it is really important for me to check this repo. Thank you in advance! The paper is great!Additional code:
app.py
file: