Open wonchul-kim opened 7 years ago
I also run it by ~.py in terminal. But, I got similar error
(10000, 1.7, 1) F tensorflow/stream_executor/cuda/cuda_dnn.cc:222] Check failed: s.ok() could not find cudnnCreate in cudnn DSO; dlerror: /usr/local/lib/python2.7/dist-packages/tensorflow/python/_pywrap_tensorflow.so: undefined symbol: cudnnCreate Aborted (core dumped)
..... Could you please help me?
@wonchul-kim Did you install cudnn? If you run gpu version, you should be installed cudnn.
If you already installed cudnn, you check path, ld_library_path, cuda_home path in the bashrc file
"export PATH=/usr/local/cuda/bin:$PATH export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64" export CUDA_HOME=/usr/local/cuda"
I've already checked the path... But, it is fine as you wrote.
When I run other reinforcement learning code,,, there is no problem... I really don't know why....
2017-05-31 13:46 GMT+09:00 ishuca notifications@github.com:
@wonchul-kim https://github.com/wonchul-kim Did you install cudnn? If you run gpu version, you should be installed cudnn.
If you already installed cudnn, you check path, ld_library_path, cuda_home path in the bashrc file
"export PATH=/usr/local/cuda/bin:$PATH export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64" export CUDA_HOME=/usr/local/cuda"
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ishuca/Reinforcement-Learning/issues/1#issuecomment-305081618, or mute the thread https://github.com/notifications/unsubscribe-auth/AT7jwV4xIxB09GrK3QvcdIfVx3w8evImks5r_PCggaJpZM4Nq6rG .
@wonchul-kim https://github.com/tensorflow/tensorflow/issues/722
You may be have to changed cudnn version 6.0 to 5.1.
Thank you! I solved the problem!!
I want to ask you that there is no plan to post another lecture related to RL? I've learned a lot about RL so that I can implement DDPG.
However, I still don't fully understand RL mathematically... Is there any reference that will be helpful for me to understand RL?
What I am working on is that I want to control the robot arm ( 6 dof, universal robot, ur3) via end-to-end method. So there is a camera above the robot arm and it captures fully robot arm frame by frame. Through the images(frames), the robot learn how to move to approach the given goal position. Now I am considering DDPG,,, but it takes so much time to converge even in the simulation....
So I am researching some papers to reduce the computational load and.... thinking .....
Do you have any advice...?
Thank you. Wonchul Kim
2017-06-01 10:32 GMT+09:00 ishuca notifications@github.com:
@wonchul-kim https://github.com/wonchul-kim tensorflow/tensorflow#722 https://github.com/tensorflow/tensorflow/issues/722
You may be have to changed cudnn version 6.0 to 5.1.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ishuca/Reinforcement-Learning/issues/1#issuecomment-305363859, or mute the thread https://github.com/notifications/unsubscribe-auth/AT7jwS0LYpc5SGjiusEtSbbOWrSKMrOnks5r_hSngaJpZM4Nq6rG .
@wonchul-kim It's not my lecture. I just study 'https://github.com/awjuliani/DeepRL-Agents'
You may be read his blog 'https://medium.com/emergent-future/simple-reinforcement-learning-with-tensorflow-part-0-q-learning-with-tables-and-neural-networks-d195264329d0'
In robot arm research, Did you try 'A3C'?
It was faster, simpler, more robust, and able to achieve much better scores on the standard battery of Deep RL tasks.
I recommend his blog 'https://medium.com/emergent-future/simple-reinforcement-learning-with-tensorflow-part-8-asynchronous-actor-critic-agents-a3c-c88f72a5e9f2'
I've got this error,
The kernel appears to have died. It will restart automatically.
Could you help me out??