Open bolshoytoster opened 5 months ago
I'm trying to port ppo2 to rust, and I've managed to mostly do this. I have, however come across something I don't understand.
In baselines/common/distributions.py, when creating the DiagGaussianDistribution, the tensor mean is multiplied by 0, then added to logstd: https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/distributions.py#L105
baselines/common/distributions.py
DiagGaussianDistribution
mean
logstd
Can anyone explain why this isn't just this?
pdparam = tf.concat([mean, logstd], axis=1)
I'm trying to port ppo2 to rust, and I've managed to mostly do this. I have, however come across something I don't understand.
In
baselines/common/distributions.py, when creating theDiagGaussianDistribution, the tensormeanis multiplied by 0, then added tologstd: https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/distributions.py#L105Can anyone explain why this isn't just this?