search

MOCR / DDPG

reimplementation of the ddpg algorithm using tensorflow
38 stars 13 forks source link

DDPG Actor output saturate #3

Open m5823779 opened 5 years ago

m5823779 commented 5 years ago

Hello~ I have some question about DDPG Ｗhen my action dimension = 1, the result is good, but when my action dimension = 2 (the activation function is tanh and sigmoid), the output of actor will saturate. Here is the result what I said: https://github.com/m5823779/DDPG By the way, I use batch normalization only in my actor network. Do you know where is the problem?

osigaud commented 5 years ago

Hi,

This repo is old stuff (3 years without a commit). Things evolved a lot in deep RL meanwhile. Check OpenAI's website for the deep RL baselines, or spinning up. You'll find state-of-the-art algorithms with clean code. But this does not mean your DDPG won't saturate anymore...