issues
search
RasmusBrostroem
/
ConnectFourRL
0
stars
0
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Update docstrings in `players.py`
#113
jbirkesteen
opened
11 months ago
0
Clean connect_four
#112
jbirkesteen
closed
11 months ago
0
Mention benchmarking and `TDAgent` in `players.py`
#111
jbirkesteen
opened
11 months ago
0
Move definition of `calculate_rewards()`
#110
jbirkesteen
closed
11 months ago
0
New training script
#109
jbirkesteen
opened
11 months ago
0
Make TDAgent hyperparameters configurable
#108
jbirkesteen
closed
11 months ago
1
Epsilon-greedy action selection
#107
jbirkesteen
closed
11 months ago
0
Use any reward for update
#106
jbirkesteen
closed
11 months ago
0
Add alpha as kwarg
#105
jbirkesteen
closed
11 months ago
1
Elaborate the update rule
#104
jbirkesteen
opened
11 months ago
2
Test training strategies for `TDAgent` and report results
#103
jbirkesteen
opened
11 months ago
2
Look closer at value estimates
#102
jbirkesteen
opened
11 months ago
3
Improve the training scripts
#101
jbirkesteen
opened
11 months ago
0
Enable use of other rewards for `TDAgent`
#100
jbirkesteen
closed
11 months ago
1
Implement epsilon-greedy `TDAgent`
#99
jbirkesteen
closed
11 months ago
0
Fix tdagent update
#98
RasmusBrostroem
closed
11 months ago
0
Let `incremental_update()` take any reward
#97
jbirkesteen
closed
11 months ago
0
Control rendering with buttons
#96
jbirkesteen
opened
11 months ago
0
Show benchmarking games in rendering
#95
jbirkesteen
opened
11 months ago
0
Show who is playing
#94
jbirkesteen
opened
11 months ago
0
Visualise probabilities
#93
jbirkesteen
opened
11 months ago
0
Double-check winrate calculation
#92
jbirkesteen
opened
11 months ago
0
Use correct states for `incremental_update()`
#91
jbirkesteen
closed
11 months ago
0
74 let humanplayer interact with mouse input
#90
RasmusBrostroem
closed
11 months ago
0
Add custom step sizes for neptune logs
#89
jbirkesteen
opened
11 months ago
0
Wrong specification of neptune run types
#88
jbirkesteen
opened
11 months ago
0
Benchmark and training attribute
#87
jbirkesteen
closed
11 months ago
0
Resetting of eligibility traces
#86
jbirkesteen
closed
11 months ago
1
Implement td agent
#85
jbirkesteen
closed
11 months ago
0
Use `training` instead of `is_training` for the `TDAgent`
#84
jbirkesteen
closed
11 months ago
1
`batch_size` and `n_updates` are hardcoded in `training_script.py`
#83
jbirkesteen
closed
11 months ago
1
Refactor use game instance
#82
RasmusBrostroem
closed
11 months ago
0
TD-backgammon inspired agent
#81
jbirkesteen
opened
1 year ago
1
Provide game object instead of just the board to players
#80
jbirkesteen
closed
11 months ago
2
`calculate_rewards()` is defined the wrong place
#79
jbirkesteen
closed
11 months ago
2
Clean players module
#78
jbirkesteen
closed
1 year ago
0
Fix requirements for no cuda
#77
jbirkesteen
closed
1 year ago
0
Incorrect reward calculation when `not_ended_reward` isn't default
#76
jbirkesteen
opened
1 year ago
0
Benchmarking during training
#75
jbirkesteen
closed
11 months ago
0
Let HumanPlayer interact with mouse input
#74
jbirkesteen
closed
11 months ago
1
Update env and initial cleaning of repo
#73
jbirkesteen
closed
1 year ago
0
Neptune import
#72
jbirkesteen
closed
1 year ago
0
Change token approach for neptune
#71
jbirkesteen
closed
1 year ago
0
Refactor the project
#70
jbirkesteen
closed
1 year ago
0
Save and load agents
#68
jbirkesteen
closed
1 year ago
2
New `select_action` failed when illegal moves not allowed
#67
jbirkesteen
opened
1 year ago
2
Create template for training scripts
#66
jbirkesteen
closed
1 year ago
1
Logging of stats
#65
RasmusBrostroem
closed
1 year ago
0
Create requirements
#64
jbirkesteen
closed
1 year ago
0
Mini max
#63
RasmusBrostroem
closed
1 year ago
0
Next