-
Hi,
Currently, I have used 4 algorithms from stable-baselines for the task of Roboschool HumanoidFlagrunHarder. My evaluation metric is the mean reward of 100 episodes. Basically: PPO2 is perfect, A2…
-
### Describe the bug
While starting a sweep through `.yml` file, the agent did not correctly started the sweep with the right combinations of parameters. Instead some hyperparameter combo is repeated…
-
Hello,
I'm having trouble loading your released hierarchical model checkpoint-- it seems to be trying to access a class from `pytorch-a2c-ppo-acktr` that doesn't exist.
```
>>> import torch
>>…
-
1 개의 액터러너를 가지고 샘플을 모아서 학습시키는 것은 학습 속도가 느린 것 같습니다. 또한 여러개의 액터러너로 학습시킨 에이전트보다 policy의 quality가 상당히 낮기 때문에 여러 개의 액터러너를 가지고 학습해야할 것 같습니다. 다음과 같은 순서로 진행하면 될 것 같습니다.
1. 여러개의 액터러너가 있는 환경 만들기
2. 각 액터러너로 …
-
Hello! Firstly, I wanted to say this mod is incredible! It truly is the best of both worlds between the game's original atmosphere and the power of distant land/long render distances.
However, whil…
-
-
@pzhokhov Is it possible to add other algorithms on benchmark Mujoco please(seems Ant environment is missing)? And It would be good to also release the data that is used to create these graphs so we c…
-
Would it be possible to make the A2C code (and other algorithms) a bit more pedagogical?
-
Hi,
Support for MultiDiscrete and MultiContinuous action spaces is not built for A2C implementation.
Does anyone know which algorithm can help me with large discrete action spaces?
Thanks!
-
Dear Author,
When I set num_processes=1, some errors occur, i could be a bug.
`File "/home/ken/project/pytorch-a2c-ppo-acktr-gail/a2c_ppo_acktr/envs.py", line 232, in step_wait
self.stacked…