-
submission-v3na
normalize adv by rollouts
-
submission-3a
recalculate advantages
-
submission-h3000
gamma 9993
-
submission-v3.32x
fs2 no rnorm
-
submission-v3.31x
fs2 clip25
-
submission-v3.3x
no fs best rew
-
submission-v3.x
best model deque100
-
submission-v3.2a
gray framestack 2
-
submission-v3.2
layernorm
-
submission-v3.1d
rnorm fix
-
submission-v3.1c
no framestack
-
submission-v3.1b
reward norm
-
submission-v3.1a
coeff bugfix
-
submission-v3.1
ray rollouts
-
submission-v3
custom torch
-
submission-v2
torch baseline
-
submission-v1.8b
lr 5e-4
-
submission-v1.8a
ds2x-modelx2-fs2-2nd-layer