werner-duvaud/muzero-general issues and pull requests

#232 - Bump actions/download-artifact from 2 to 4.1.7 in /.github/workflows

Pull Request - State: open - Opened by dependabot[bot] 18 days ago
Labels: dependencies

#231 - Chess and other non-trivial games

Issue - State: open - Opened by StepHaze 7 months ago - 2 comments
Labels: enhancement

#230 - Update game logic and error handling in spiel.py

Pull Request - State: open - Opened by leoni-q 9 months ago

#229 - cross play between models

Issue - State: open - Opened by dmtrung14 9 months ago
Labels: enhancement

#228 - What does the `replay_buffer.pkl` do?

Issue - State: closed - Opened by dmtrung14 10 months ago - 1 comment

#227 - How can I use a pre-trained model?

Issue - State: open - Opened by worldsoft 12 months ago - 1 comment
Labels: enhancement

#226 - MuZero choose the same action

Issue - State: open - Opened by sdumi03 about 1 year ago
Labels: bug

#225 - Muzero crashes when choosing spiel/backgammon

Issue - State: open - Opened by artshar about 1 year ago
Labels: bug

#224 - Fix: gym v26 migration

Pull Request - State: open - Opened by lduchosal about 1 year ago

#223 - Breakout: ModuleNotFoundError shows gym[atari] instead of python-opencv

Issue - State: open - Opened by velicanerdem over 1 year ago - 1 comment
Labels: bug

#222 - question about action encoded

Issue - State: open - Opened by Nightbringers over 1 year ago
Labels: bug

#221 - The difference between offical pseudo code and this repository about "num_unroll_steps"

Issue - State: open - Opened by ZF4444 over 1 year ago
Labels: bug

#220 - Dirichlet noise added outside of training

Issue - State: open - Opened by TommyX12 over 1 year ago
Labels: bug

#219 - Update gym package to gymnasium

Issue - State: open - Opened by Mlokos over 1 year ago - 1 comment
Labels: enhancement

#218 - fix: fixed ray's error 'No module named aiohttp.signals'

Pull Request - State: open - Opened by ChunchangShao over 1 year ago

#217 - Without Selfplay, why 1 games on a running??

Issue - State: open - Opened by dlrlfkr11 over 1 year ago - 2 comments
Labels: enhancement

#216 - [WIP] Sampled Muzero

Pull Request - State: open - Opened by JosephDenman almost 2 years ago - 1 comment

#215 - Mean_value plot in Total_reward - Interpretation

Issue - State: open - Opened by SunilaAkbar almost 2 years ago

#214 - Question about the dimension of value and reward network

Issue - State: open - Opened by jiachengc almost 2 years ago
Labels: bug

#213 - Can't train using GPU? The torch version for this environment is '1.10.0cpu', that is, CPU one.

Issue - State: open - Opened by SunilaAkbar almost 2 years ago - 2 comments
Labels: enhancement

#212 - Question about the perspective transformation of two players when calculating Q?

Issue - State: open - Opened by puyuan1996 almost 2 years ago

#211 - The model does not converge for breakout

Issue - State: open - Opened by yungangwu almost 2 years ago - 13 comments
Labels: enhancement

#210 - Only One Player: Can we use MuZero?

Issue - State: open - Opened by 1121091694 almost 2 years ago - 2 comments
Labels: enhancement

#209 - Switch architecture to parallel sync ray tasks

Pull Request - State: closed - Opened by cmarlin almost 2 years ago

#208 - TypeError: can't pickle function objects

Issue - State: open - Opened by OopsYouDiedE almost 2 years ago
Labels: bug

#207 - Question: Does muzero-general support 2 player games with simultaneous action selection?

Issue - State: open - Opened by moscoso about 2 years ago - 3 comments

#206 - Uncertainty pls

Pull Request - State: closed - Opened by Dirichi about 2 years ago - 1 comment

#205 - Small ensemble

Pull Request - State: closed - Opened by Dirichi about 2 years ago - 1 comment

#204 - Consistency

Pull Request - State: closed - Opened by Dirichi about 2 years ago

#203 - Remove Batch Norm?

Issue - State: open - Opened by verbose-void about 2 years ago
Labels: enhancement

#202 - OpenGL rendering on a remote server over X11

Issue - State: open - Opened by jrjbertram about 2 years ago
Labels: bug

#201 - [Question] Is the environment required to have no hidden information?

Issue - State: closed - Opened by ZhengWenZhang about 2 years ago

#200 - sampling in continuous/complex action spaces with 'density prior' is not working

Issue - State: open - Opened by ManorZ about 2 years ago
Labels: bug

#199 - Batch MCTS

Issue - State: open - Opened by szrlee about 2 years ago
Labels: enhancement

#198 - raw install has ray problem

Issue - State: open - Opened by EngrStudent about 2 years ago - 1 comment
Labels: bug

#197 - Render model

Issue - State: open - Opened by theeduardomora about 2 years ago - 1 comment
Labels: bug

#196 - Is there a place we can see and share results for each game?

Issue - State: open - Opened by onegigbyte over 2 years ago

#195 - custom observation transformation

Issue - State: open - Opened by SimpleMathmatics over 2 years ago

#194 - Fix requirements in ci

Pull Request - State: closed - Opened by werner-duvaud over 2 years ago
Labels: bug

#193 - Policy target after MCTS should be in form of probabilities

Issue - State: open - Opened by 2M-kotb over 2 years ago - 1 comment

#192 - my first commit

Pull Request - State: closed - Opened by aliigii over 2 years ago

#191 - Sampled MuZero implementation

Issue - State: open - Opened by matthiaskiller over 2 years ago - 1 comment
Labels: enhancement

#190 - fix a bug due to the use of DataParallel

Pull Request - State: closed - Opened by vincentzhang over 2 years ago

#189 - Can muzero learn to play two different games at the same time

Issue - State: open - Opened by lwaif over 2 years ago - 1 comment
Labels: enhancement

#188 - Add workflow

Pull Request - State: closed - Opened by ahainaut over 2 years ago

#187 - Add github workflow

Pull Request - State: closed - Opened by ahainaut over 2 years ago

#186 - Strange observations and actions in continuous implementation.

Issue - State: closed - Opened by dylanamiller over 2 years ago - 1 comment

#185 - MuZero Unplugged

Issue - State: open - Opened by tbskrpmnns over 2 years ago - 7 comments
Labels: enhancement, question

#184 - Why is root.visit_count initialized to 0 and root_predicted_value not included in root node value?

Issue - State: open - Opened by dniku over 2 years ago
Labels: enhancement

#183 - procgen

Issue - State: open - Opened by hlsfin over 2 years ago
Labels: enhancement, question

#182 - ray.exceptions.RayActorError: The actor died unexpectedly before finishing this task.

Issue - State: open - Opened by hairinwind over 2 years ago - 1 comment

#181 - Why my game does not remember the steps trained

Issue - State: closed - Opened by hairinwind over 2 years ago - 3 comments

#180 - training result cannot be loaded on another machine

Issue - State: open - Opened by hairinwind over 2 years ago - 1 comment

#178 - Scaling of historical stacked observations

Issue - State: closed - Opened by tuero almost 3 years ago - 1 comment

#177 - added support for multiple dimension continuous action spaces

Pull Request - State: open - Opened by devin-m-NRL almost 3 years ago - 1 comment

#176 - Total Training Reward rises then drops again

Issue - State: closed - Opened by annahambi almost 3 years ago - 2 comments
Labels: question

#175 - Entropy loss in continuous actions

Issue - State: closed - Opened by 2M-kotb almost 3 years ago - 4 comments

#174 - RuntimeError: module must have its parameters and buffers on device cuda:0 (device_ids[0]) but found one of them on device: cpu

Issue - State: open - Opened by lukaszkn almost 3 years ago - 3 comments

#173 - Is there a more optimized way to complete training? Every game has to be learned for so long, people learn to master it early

Issue - State: closed - Opened by lwaif about 3 years ago - 1 comment
Labels: question

#172 - Faster calculations in self_play.py

Pull Request - State: closed - Opened by bibidybop about 3 years ago

#171 - Struggling to get Ray working

Issue - State: open - Opened by SheldonCurtiss about 3 years ago
Labels: question

#170 - Target Value Offset

Issue - State: closed - Opened by dans-acc about 3 years ago - 1 comment

#169 - Optimization of some parameters for tictactoe.

Pull Request - State: open - Opened by AdrianAcala about 3 years ago - 4 comments

#168 - Add progress bar with speed and estimation

Pull Request - State: closed - Opened by LeoVS09 about 3 years ago

#167 - Scrabble implementation - How to include Player's rack observation

Issue - State: closed - Opened by nicolasnijssen about 3 years ago - 2 comments

#166 - Would it possible to write go game and chess game program?

Issue - State: closed - Opened by leqingli2000 about 3 years ago - 1 comment
Labels: enhancement

#165 - Dimensionality issue in continuous action space

Issue - State: closed - Opened by alik604 about 3 years ago - 1 comment

#164 - could it run without ray?

Issue - State: closed - Opened by hilberthu over 3 years ago - 2 comments
Labels: question

#163 - File not found ray distributed cluster worker node save checkpoint

Issue - State: open - Opened by saintpoida over 3 years ago - 3 comments

#162 - If I know the environment, is it better to train alphazero?

Issue - State: open - Opened by omgmax over 3 years ago - 1 comment
Labels: question

#161 - 2 players moving simultaneously

Issue - State: open - Opened by omgmax over 3 years ago - 2 comments
Labels: question

#160 - Adapting to RLLib

Issue - State: closed - Opened by SebastianBodza over 3 years ago - 1 comment
Labels: enhancement

#159 - Self-play very slow and inefficient on GPU (self.selfplay_on_gpu = True)

Issue - State: open - Opened by kevaday over 3 years ago - 1 comment
Labels: enhancement

#158 - How to adapt Muzero to financial trading?

Issue - State: closed - Opened by Ray-0403 over 3 years ago - 21 comments

#157 - RAM memory usage of SelfPlay.continuous_self_play() keeps growing

Issue - State: open - Opened by adalsteinnpals over 3 years ago - 8 comments
Labels: help wanted

#156 - Input of representation function is 131 planes instead of 128 !!

Issue - State: closed - Opened by 2M-kotb over 3 years ago

#155 - Training step always 0

Issue - State: open - Opened by yeekit24 over 3 years ago - 2 comments

#153 - Reward computation based on next state instead of (state, action)

Issue - State: closed - Opened by FXDevailly over 3 years ago - 3 comments
Labels: question

#152 - Interpreting the training curves on Tic Tac Toe

Issue - State: closed - Opened by itmorn over 3 years ago - 13 comments
Labels: documentation

#151 - Keep replay buffer on disk (not in memory), allowing it to grow to any size.

Pull Request - State: open - Opened by me-unsolicited over 3 years ago - 2 comments

#150 - [3090 rtx] Very slow training on resnet with 1 block

Issue - State: closed - Opened by HadiSDev over 3 years ago - 5 comments

#149 - Render history

Pull Request - State: open - Opened by egafni over 3 years ago

#144 - Why did I train lunarlander so slowly？more than 1 hour used 1 gpu

Issue - State: closed - Opened by Augustiu over 3 years ago - 1 comment
Labels: question

#143 - Alpha Zero / MuZero differences

Issue - State: closed - Opened by Counterfeiter over 3 years ago - 5 comments
Labels: enhancement, question

#142 - Updates on Reanalyse / Sample Efficiency (Re-executing MCTS, Parallelization, Stabilization with a target model, etc.)

Pull Request - State: open - Opened by FXDevailly over 3 years ago - 6 comments

#141 - Only single step two player mode, no real two player mode!?

Issue - State: closed - Opened by Counterfeiter over 3 years ago - 2 comments
Labels: enhancement, question

#139 - RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED

Issue - State: closed - Opened by theword over 3 years ago - 10 comments

#138 - Reason for not re-executing MCTS and updating policy "targets" (child visits) in "Reanalyze" ?

Issue - State: closed - Opened by FXDevailly over 3 years ago - 9 comments
Labels: enhancement

#135 - Can training continue from a previous trained model?

Issue - State: closed - Opened by atalapan over 3 years ago - 4 comments

#132 - Using TPUs

Issue - State: open - Opened by StrangeTcy over 3 years ago - 1 comment
Labels: enhancement

#130 - POMDP

Issue - State: closed - Opened by WenyuHan-LiNa over 3 years ago - 5 comments
Labels: question

#127 - Improving Atari hyperparameters

Issue - State: open - Opened by xiaolonghao over 3 years ago - 4 comments
Labels: question

#103 - Add open_spiel game wrapper

Pull Request - State: closed - Opened by goshawk22 over 3 years ago - 2 comments

#100 - Typo In muzero.py comments

Issue - State: closed - Opened by FrancescoVassalli over 3 years ago - 1 comment

GitHub / werner-duvaud/muzero-general issues and pull requests