-
When I set env.seed(0) (or some other seed) I expected all random elements of env to produce deterministically. However, the env.action_space.sample() function still seems to output randomly.
```
…
-
### What is the problem?
I've been using JAX as my framework for a little while now. I just upgraded to the nightly build (due to some unrelated issues) and now RLlib is telling me I need to instal…
-
-
From what I understand, `env.step` returns `observation, reward, done, info`. `done` is supposed to indicate whether the agent reached the goal or fell into a hole (terminal states). But sometimes, it…
-
Hi,
First, Great Work on TF Agents !! :)
I encountered a problem on running train_eval and also the dqn_tutorial colab when changing the environment to FrozenLake .
The code that I tried :
```
e…
-
Hi everyone, I meet a problem in the render of the frozen lake. Actually I was testing it and choose the move of the cursor by myself, not random and not even chose by an agent.
The problem is th…
-
Based on the linked article below, the reward value at each time step should be +1.0 for reaching the goal, -0.2 for agent death, and -0.01 for reaching a non-goal frozen spot. What seems to be happen…
-
I am just trying to start up in learning how this stuff works, but im immediately getting met by an error from the frozenlake stuff. I have no idea why it's happening or what I can do to fix it? Pleas…
-
We should write a more detailed explanation of every environment, in particular, how the reward function is computed.
-
After run FrozenLake_v0.py, the agent didn't learn anything. I'm wondering which part is wrong.