search

Kyziridis / BipedalWalker-v2

Solving openAI's game 'BipedalWalker-v2' with Deep Reinforcement Learning
GNU General Public License v2.0
26 stars 6 forks source link

Question about rl-Bipedal #4

Open GHysmu opened 4 years ago

GHysmu commented 4 years ago

I try to run the code rl-Bipedal but something wrong with it. I found the reason is that in the class PPO, the defination of loss and some other variables(including the initialization) are not in any function. All these code never been used. I don't dnow if I change something wrong or any other problem. Thank you!

GHysmu commented 4 years ago

Besides,there may be something wrong with "f.write('\n' + str(np.mean(rewards_over_time[-100:])))" ,may be you have to change the "rewards_over_time" to"rew_list" which you defined before.

Kyziridis commented 4 years ago

Hi bella, I fixed the bugs in the rl-bipedal.py script. I think that now is running correctly. Thank you for your interest in my repo.

On Tue, 26 May 2020 at 10:56, bella notifications@github.com wrote:

Besides,there may be something wrong with "f.write('\n' + str(np.mean(rewards_over_time[-100:])))" ,may be you have to change the "rewards_over_time" to"rew_list" which you defined before.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/Kyziridis/BipedalWalker-v2/issues/4#issuecomment-633899036, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE2LEFKWCDY6CPDI5SVIZVDRTN74FANCNFSM4NKDLDFQ .

-- Giorgos

Kyziridis commented 1 year ago

Fixed >_