huggingface/deep-rl-class

[HANDS-ON BUG] Unit#6 NameNotFound: Environment 'AntBulletEnv' doesn't exist.

ShridiptaSatpati opened this issue · 9 comments

ShridiptaSatpati commented

Hey, it seems I'm getting the following error while trying execute the function make_vec_env with env_id="AntBulletEnv-v0" environment:

NameNotFound: Environment 'AntBulletEnv' doesn't exist.

but gym creates the "AntBulletEnv-v0" in one of the previous cells of the Unit 6 notebook without any problem.

snapshot:
Screenshot (203)

I am using Google Colab

thanks in advance for your assistance

arminmrm commented

Hi! I exactly experienced this issue. I was able to fix it as follows:
env = make_vec_env(lambda: gym.make(env_id), n_envs=4)

  • 🎉1
ShridiptaSatpati commented

The training process executed without any hindrance after modifying according to your suggested changes but encountering the error AssertionError: The render_mode must be 'rgb_array', not None while pushing the trained agent to the hub.

image

  • 😕1
arminmrm commented

Yes, I exactly encountered the same issue, as well. I opened an issue for it and waiting for a response.

  • 👍1
arminmrm commented

Hey @ShridiptaSatpati! Did you figure this out?

ShridiptaSatpati commented

@arminmrm Unfortunately I could not fix the problem. But Thomas Simonini and his team promised that they are still updating the Unit 6 notebook and they would be notifying us when it is ready.

image

You can join the discord server to receive the latest updates.
https://huggingface.co/learn/deep-rl-course/unit0/discord101?fw=pt has that discord link.

  • 👍3
MattBoraske commented

+1 for encountering this issue as well.

taleslimaf commented

+1 for encountering this issue as well.

MateoCamara commented

+1 for encountering this issue as well.

dummy-round commented

Hi @ShridiptaSatpati, I think the problem is from the Stable-baselines3 new release: https://github.com/DLR-RM/stable-baselines3/releases/tag/v2.0.0 and they switch to Gymnasium for the backend, and Gymnasium seems not support the AntBulletEnv. So workaround is to try to install an older release of stable-baselines3. I added a line in the Colab: !pip install stable-baselines3[extra]==1.8.0 and it worked.

  • 👍5