An image processing application & library built in C++20 and the Qt Framework.
A pipeline using big data technologies
BNN is a library for binary neural network. In which we will introduce state of the art algorithms and approaches.
Competitive Programming Library
A Platform for sharing computer science knowledge and improving problem solving
My attempt to explore Machine Learning & Data Science
ramizouari/ArtificialIntelligence 2
This repository contains a list of academic problems that we solved using classical AI.
a little chatroom
push eventramizouari/StochasticGames
commit sha 383c067519df9e1cbf0f54ad09109107c43693b8
Updating introduction
commit sha f76315bebaa608261f11c62c32716a9590282f05
Adding introductions and conclusions
push time in 21 days
push eventramizouari/StochasticGames
commit sha a5fee8d536b5aa496894bd7c2384cbd17b665a10
Chapter RL almost completed
commit sha ee404a550d0698921b276fb614eae85179d17978
Updating format to INSAT standard
push time in 23 days
push eventramizouari/StochasticGames
commit sha 129c6340398ddafeebc09d7d19fae4502cf1f72a
Adding acronym lists
push time in 24 days
push eventramizouari/StochasticGames
commit sha e55179b708c6ea43cf4fb8a3b3dc122af12330a9
Adding RL/SP chapter
push time in 25 days
push eventramizouari/StochasticGames
commit sha e0ce2fc26e398e747355e24a83bb12874125abc4
Adding Pipeline ardiagram
push time in a month
push eventramizouari/StochasticGames
commit sha 1979f22b44017a4531c69c7aa673ae0d611e1f0c
Refining proofs + A Adding class diagrams
push time in a month
push eventramizouari/StochasticGames
commit sha 7be204d14499567be7bb5a38de4b09bd2a59ec57
Major update to report
push time in a month
startedfacebookresearch/audiocraft
started time in 2 months
push eventramizouari/StochasticGames
commit sha 53ae1afa29680691b38529b622b81e01f5cc7d29
Updating open_spiel
push time in 3 months
push eventramizouari/open_spiel
commit sha b33d31714b174a4ea78440944a41275d922a1813
Fixing Evaluator save thread
push time in 3 months
push eventramizouari/StochasticGames
commit sha 85d1d0b059604c97e5e4cabda7ab2c98798856a1
Updating Report
push time in 3 months
push eventramizouari/StochasticGames
commit sha 7fd0832f66fd74d81162e37acaa189c9c07b7364
Adding report
push time in 3 months
push eventramizouari/open_spiel
commit sha 9c4d37cc1747ed948346b4d5871dd7248aeb871a
More robust RNG + Support for reseeding
push time in 3 months
push eventramizouari/StochasticGames
commit sha 9757fe57f6bcdba357c23411ee5e524ad51859a6
Updating open_spiel
push time in 3 months
push eventramizouari/open_spiel
commit sha 55684f1c4ecaa569c9ba31dc7150b4c75c9faf07
Adding payoff noise + view Replay Buffer infos
push time in 3 months
push eventramizouari/StochasticGames
commit sha 9f9f2dc5a85a0dfc6cc9f095d4c34de89c756b85
Update open_spiel
push time in 3 months
push eventramizouari/open_spiel
commit sha e41976de8b56452ff8d64e0af6026076d2060510
Adding service type to config
push time in 3 months
push eventramizouari/StochasticGames
commit sha 35dfd9ffa0b1913bb5ea62b4b6105284d090b324
Updating open_spiel
push time in 3 months
push eventramizouari/open_spiel
commit sha 38b0737d652c31f5d5236e9bd2022d7542e5b905
Adding template configuration
push time in 3 months
issue commentdeepmind/open_spiel
Hello @tewalds.
Thank you for your reply.
In the game that I am working on (Mean Payoff Game), I had to deploy it on a HPC cluster for faster trajectory generation. For my use case, trajectories were sent with reverb, and model broadcasting the model and monitoring were done with a HTTP server on each service.
Also, the HTTP part is generic on the sense that we can switch it with another protocol (or default to the multiprocessing queues as is implemented by default), one only have to change:
- Model broadcasting function
- The ReplayBuffer implementation (Local via queues / Reverb via gRPC, or a custom one)
- The model update part on the actors and evaluators.
I had to switch to Python on that part due to the lack of documentation of C++'s implementation of Reverb, but of course it is doable. I will need to contact the Reverb team for more intuition on their C++ code.
Also another limitation of the C++ part, I am still not find the correct format to call the fit function on C++. I only was able to do inference. Also I was not able to load individual checkpoints, but that can be mitigated by simply loading the whole SavedModel bundle on each update.
On the other hand, assuming the Reverb problem in C++ will be resolved, what I can do is implement the learner in Python, and the actors and evaluators in C++. And have them communicate using for example HTTP + Reverb.
Now, as that will constitute a big code addition, I think it will be best if we split them on PR at a time. And for that I will start with the TF2 update.
And as a performance measure, can you please tell me what games should the new implementation be able to learn?
comment created time in 3 months