issues
search
zapper-95
/
Coup-RL
Models trained to play the card game Coup
0
stars
0
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Merging ctde
#51
zapper-95
closed
5 months ago
0
Choose kill card
#50
zapper-95
closed
6 months ago
1
Feature - Hyperparameter Tuning
#49
zapper-95
closed
5 months ago
0
Feature - Allow players to choose which card to kill
#48
zapper-95
closed
6 months ago
1
Random wr logging
#47
zapper-95
closed
6 months ago
0
Render mode for playing where you cannot see the agents hand
#46
zapper-95
closed
6 months ago
0
Change model to use recurrent features
#45
zapper-95
opened
6 months ago
0
Add previous actions to state
#44
zapper-95
closed
6 months ago
0
Models play one another
#43
zapper-95
closed
6 months ago
0
Feature - Allow models to be played against one another
#42
zapper-95
closed
6 months ago
0
Feature - Reshuffle on proving card
#41
zapper-95
closed
5 months ago
0
Feature - Choose card to kill
#40
zapper-95
closed
6 months ago
0
Ray rl llib
#39
zapper-95
closed
6 months ago
0
Experiment - Compare Algorithmic Approaches
#38
zapper-95
closed
5 months ago
0
Feature - Multiplayer AI better than random
#37
zapper-95
closed
5 months ago
0
Feature - Multiplayer environment implementation
#36
zapper-95
opened
6 months ago
0
Experiment - See how the number of past actions varies performance
#35
zapper-95
closed
5 months ago
0
Feature - Allow a variable number of actions to be stored in the state
#34
zapper-95
closed
6 months ago
0
Bug - Fix or supress deprecated libraries
#33
zapper-95
closed
6 months ago
1
Experiment - Compare to Starcheus' algorithms
#32
zapper-95
closed
5 months ago
0
Feature - Log winrate against random during training
#31
zapper-95
closed
6 months ago
0
Feature - Change checkpoints so they are agnostic to python version
#30
zapper-95
opened
6 months ago
0
Feature - Set up tests for desired behaviour
#29
zapper-95
closed
5 months ago
0
Ray RLlib
#28
zapper-95
closed
6 months ago
1
Test - MARLlIB
#27
zapper-95
closed
6 months ago
1
Feature - Set up tests for particular scenarios
#26
zapper-95
closed
6 months ago
1
Feature - Update action space to include previous actions
#25
zapper-95
closed
6 months ago
1
Feature - Print action distributions
#24
zapper-95
closed
6 months ago
1
Train model against previous model
#23
zapper-95
closed
6 months ago
0
Zapper 95/incorrect reward structure
#22
zapper-95
closed
6 months ago
0
Feature - Train models against previous model
#21
zapper-95
closed
6 months ago
1
Bug - Incorrect reward structure for PPO
#20
zapper-95
closed
6 months ago
0
Two players can win
#19
zapper-95
closed
7 months ago
0
Command line arguments for training testing and evaluation
#18
zapper-95
closed
8 months ago
0
Bug - Two players can win with a reward of 1
#17
zapper-95
closed
6 months ago
1
Bug - When in dead state, gives redundant options
#16
zapper-95
closed
6 months ago
1
Bug - Challenging correct counteractions seems not to work for assassination
#15
zapper-95
closed
8 months ago
0
moved readme
#14
zapper-95
closed
8 months ago
0
Reorganised directories, and cleaned some code
#13
zapper-95
closed
8 months ago
0
Reorganise Folders
#12
zapper-95
closed
8 months ago
0
Zapper 95/act co ch is buggy
#11
zapper-95
closed
8 months ago
0
Player turns seems correct if they challenge or are counteracted and …
#10
zapper-95
closed
8 months ago
0
Zapper 95/fix challenge action mask
#9
zapper-95
closed
8 months ago
0
Feature - Not skipping players turns
#8
zapper-95
closed
8 months ago
0
Bug - Check whether having the agent as player 2 effects performance
#7
zapper-95
opened
8 months ago
0
Feature - Be able to play against the best agent
#6
zapper-95
closed
8 months ago
0
Feature - Add training to be done in parrallel
#5
zapper-95
closed
5 months ago
0
Bug - Act -> Cou -> Chal seems buggy
#4
zapper-95
closed
8 months ago
0
Bug - Action mask allowed challenge as first action for an agent
#3
zapper-95
closed
8 months ago
1
Zapper 95/stop agents taking illegal actions, with action masking
#2
zapper-95
closed
8 months ago
0
Next