DDPG doesn't work in mujoco

[Kchu]

I recently tested the baselines' DDPG algorithm, but unexpectedly, DDPG failed to get great scroes and to converge. I can't get the ideal results in many mujoco environments.
I use the download source installation, the command line runned is below:
python -m baselines.run --alg=ddpg --env=HalfCheetah-v2 --num_timesteps=1e6

[ghost]

#938 This might be helpful?

[christopherhesse]

Closing in favor of #938