Yang Yue

@rosinality hello, I have the same problem "subprocess.CalledProcessError: Command '['ninja', '-v']' returned non-zero exit status 1". There is my environment: pytorch 1.3.1 torchvision 0.4.2 tensorflow 1.14 CUDA 10.0 Ubuntu 16.04...

This is my full error. @rosinality 

@rosinality, there are all my error messages. ``` Traceback (most recent call last): File "/usr/local/lib/python3.6/dist-packages/torch/utils/cpp_extension.py", line 1030, in _build_extension_module check=True) File "/usr/lib/python3.6/subprocess.py", line 438, in run output=stdout, stderr=stderr) subprocess.CalledProcessError: Command...

@rosinality I have installed python3.6-dev and python-dev. However, I still have the problem like this, ``` Traceback (most recent call last): File "/usr/local/lib/python3.6/dist-packages/torch/utils/cpp_extension.py", line 1030, in _build_extension_module check=True) File "/usr/lib/python3.6/subprocess.py",...

@rosinality I sincerely appreciate your help, saving lots of my time. I found some header files in my CUDA are missing for some reasons. I have solved the problem by...

Additionally, I use the test bash script to evaluate the model checkpoint saved in my running 1. The result is as below ``` [2022-04-06 13:47:41,957][test][INFO][main.py>] ==> Test Mean Score: 9603.125...

I have the same problem when running jax.random.split(). jax 0.3.13 jaxlib 0.3.10+cuda11.cudnn82

It seems in DMcontrol there is no true terminal state. So it allows infinte bootstrap.

For @geekyutao 's question, the point is that the next_ob will never be the start observation of the next episode. Because at the previous timestep, the next_ob is the terminal...

In my understanding, the key point is that after sampling trajectories, the agent parameters would be updated several times (it's up to args.ppo_epoch). At the first updating time, the situation...