-
I got a runtime error when trying to share a CUDA model with a child process.
The code I was running:
```
import torch
import torch.multiprocessing as mp
import torch.nn as nn
def trai…
-
There is two problems:
- When the state has a specific size (**210**), we have an error, and the actor die
- A warning message about non-writable array from Numpy (discussed previously on the slac…
-
This is not a contribution.
Versions:
python: 3.6.8
ray: 1.0
pytorch: 1.6
tensorflow: 1.15
OS: Ubuntu 18.04 Docker
Since upgrading to 0.8.7 and 1.0, we are experiencing multiple stability …
-
Hello,
I spotted some bugs in the implementation of DDPG in PyTorch.
* The gradient clipping is not implemented correctly, it uses the 'grad_norm_clipping' parameter instead of 'grad_clip' and the…
-
### System information
- **OS Platform and Distribution **: Ubuntu 16.04.2 LTS
- **Ray installed from (source or binary)**: binary
- **Ray version**: 0.5.3
- **Python version**: 3.6.7
### …
-
### System information
- **OS Platform and Distribution (Linux): amplab.cs.berkeley.edu/jenkins**:
- **Ray installed from**: source
- **Ray version**: 0.5
- **Python version**: 2.7
- **Exact …
-
I met a problem when I'm running
env = gym_super_mario_bros.make(env_id)
while my env_id = "SuperMarioBrosNoFrameskip-1-1-v0"
the complete error report is like this:
File "/home/linc/anaconda3/e…
-
Hi, I'm using VizDoom to train my agent to learn to navigate. But there are several questions:
1. in my scenario, the maze is discrete, which means I regard the vizdoom as a 3D grid world. However, …
-
### What is the problem?
PyTorch is supposed to be thread-safe, as long as you don't write a tensor using multiple threads. Please see https://discuss.pytorch.org/t/is-pytorch-supposed-to-be-th…
-
### What is the problem?
in the constructor in /usr/local/lib/python3.6/dist-packages/ray/rllib/policy/torch_policy.py there's casting of torch model to GPU if CUDA is available, regardless of a …