Commit be34b898 authored by Dipam Chakraborty's avatar Dipam Chakraborty
Browse files

remove rnorm change

parent 7de5d293
......@@ -145,7 +145,7 @@ class RewardNormalizer(object):
self.ret = self.ret * self.gamma + rews
self.ret_rms.update(self.ret)
rews = np.clip(rews / np.sqrt(self.ret_rms.var + self.epsilon), -self.cliprew, self.cliprew)
self.ret[np.array(news, dtype=bool)] = 0. ## Values should be True of False to set positional index
# self.ret[np.array(news, dtype=bool)] = 0. ## Values should be True of False to set positional index
return rews
class RunningMeanStd(object):
......
......@@ -80,7 +80,7 @@ procgen-ppo:
# nlatents: 1024
depths: [32, 64, 64]
nlatents: 512
init_normed: True
init_normed: False
use_layernorm: True
num_workers: 7
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment