Andrew Cohen
|
8ef0b3a8
|
opponent observations
|
4 年前 |
|
Andrew Cohen
|
a8f2f613
|
no energy penalty
|
4 年前 |
|
Andrew Cohen
|
1f2453ef
|
slower tennis
|
4 年前 |
|
Andrew Cohen
|
0d943676
|
inc energy pen
|
4 年前 |
|
Andrew Cohen
|
d77f2566
|
energy usage penalty to prevent superstition on serve
|
4 年前 |
|
Andrew Cohen
|
cde22a14
|
fix clipping
|
4 年前 |
|
Andrew Cohen
|
d5428487
|
addforce and static walls
|
4 年前 |
|
Andrew Cohen
|
bbc1014a
|
reduce learning rate
|
4 年前 |
|
Andrew Cohen
|
9d5d6fa7
|
Merge branch 'master' into asymm-envs
|
4 年前 |
|
Andrew Cohen
|
dab592d0
|
Merge branch 'master' into asymm-envs
|
4 年前 |
|
Andrew Cohen
|
8fba6faa
|
increase network capacity
|
4 年前 |
|
Andrew Cohen
|
843ddfc3
|
updated heuristics of cubewars
|
4 年前 |
|
Andrew Cohen
|
2c42f577
|
Merge branch 'master' into asymm-envs
|
4 年前 |
|
Andrew Cohen
|
052f1c87
|
reduce gamma
|
4 年前 |
|
Andrew Cohen
|
b80d6228
|
randomize ball/agent spawn
|
4 年前 |
|
Andrew Cohen
|
14df5d02
|
increase gamma
|
4 年前 |
|
Andrew Cohen
|
376af981
|
lower agent height
|
4 年前 |
|
Andrew Cohen
|
a56643bb
|
tennis brain to prefab
|
4 年前 |
|
Andrew Cohen
|
7da1869a
|
new tennis brain
|
4 年前 |
|
Andrew Cohen
|
e7922b68
|
trying larger beta
|
4 年前 |
|
Andrew Cohen
|
e0723381
|
new tennis brain
|
4 年前 |
|
Andrew Cohen
|
1c2e1d79
|
increase beta
|
4 年前 |
|
Andrew Cohen
|
b7bd4c2c
|
reduce winning reward
|
4 年前 |
|
Andrew Cohen
|
39e0bbe9
|
remove debug log
|
4 年前 |
|
Andrew Cohen
|
e3f6c716
|
higher granularity curr
|
4 年前 |
|
Andrew Cohen
|
4769cb1e
|
proximity bonus
|
4 年前 |
|
Andrew Cohen
|
1b10ef6d
|
clip speed
|
4 年前 |
|
Andrew Cohen
|
b0243014
|
broken prefab..
|
4 年前 |
|
Andrew Cohen
|
251dcc76
|
remove timepenalty from tennis
|
4 年前 |
|
Andrew Cohen
|
fb7aa862
|
remove timepenalty obs
|
4 年前 |
|
Andrew Cohen
|
b80c48bf
|
new cube wars brains
|
4 年前 |
|
Andrew Cohen
|
fda39c3d
|
more beta tuning...
|
4 年前 |
|
Andrew Cohen
|
a5ca5e0c
|
reduce beta for new reward func
|
4 年前 |
|
Andrew Cohen
|
a6e6e63e
|
timestep penalty on loss only
|
4 年前 |
|
Andrew Cohen
|
547f3192
|
beta .05
|
4 年前 |
|
Andrew Cohen
|
5d22b819
|
added timepenalty to obs
|
4 年前 |
|
Andrew Cohen
|
0871fc96
|
remove beta/no curr tennis
|
4 年前 |
|
Andrew Cohen
|
54972202
|
tuning beta tennis
|
4 年前 |
|
Andrew Cohen
|
3f806353
|
increased beta
|
4 年前 |
|
Andrew Cohen
|
e9f570aa
|
slightly larger beta tennis
|
4 年前 |
|
Andrew Cohen
|
443935f4
|
remove time bonus loss
|
4 年前 |
|
Andrew Cohen
|
a1143427
|
increased entro bonus tennis
|
4 年前 |
|
Andrew Cohen
|
428b1dfa
|
Hit bonus for whole exp
|
4 年前 |
|
Andrew Cohen
|
1c4ba1a5
|
add timestep bonus to loss
|
4 年前 |
|
Andrew Cohen
|
717fae65
|
reduce tennis latest_model_ratio
|
4 年前 |
|
Andrew Cohen
|
3df4f4a3
|
smaller window cubewar
|
4 年前 |
|
Andrew Cohen
|
32f562d9
|
striker goalie increase latest_mod ratio
|
4 年前 |
|
Andrew Cohen
|
ca6cdff3
|
fixed broken prefab...
|
4 年前 |
|
Andrew Cohen
|
028a8d59
|
larger network/6 stacked obs
|
4 年前 |
|
Andrew Cohen
|
d54fdfbf
|
increase batch/buff/erbeta
|
4 年前 |
