Views
No views yet
| Run | Mean Reward | Std Dev |
|---|---|---|
| Seed 0 | -0.15 | 0.28 |
| Seed 1 | -0.40 | 0.41 |
| Seed 2 | -0.12 | 0.30 |
1from stable_baselines3 import PPO
2model = PPO.load("0/ckpt_latest.zip")fluidgym==0.0.2. In order to use
them with newer versions of FluidGym, you need to wrap the environment with a
FlattenObservation wrapper as shown below:1import fluidgym
2from fluidgym.wrappers import FlattenObservation
3from stable_baselines3 import PPO
4
5env = fluidgym.make("CylinderJet3D-medium-v0")
6env = FlattenObservation(env)
7model = PPO.load("path_to_model/ckpt_latest.zip")
8
9obs, info = env.reset(seed=42)
10
11action, _ = model.predict(obs, deterministic=True)
12obs, reward, terminated, truncated, info = env.step(action)