openai / gym

A toolkit for developing and comparing reinforcement learning algorithms.
https://www.gymlibrary.dev
Other
34.73k stars 8.61k forks source link

HandReach-v0 goal position not visible in video #1594

Closed chaitanyadwivedii closed 2 years ago

chaitanyadwivedii commented 5 years ago

After running the following python3 -m baselines.run --alg her --env HandReach-v0 --num_timesteps 0 --load_path model/t1 --play

I get the following output (attached) Screenshot 2019-07-09 at 1 00 24 PM As can be seen the goal positions are not visible in the image. Any reason this might be happening ? Or is this the expected behaviour? I was expecting something like this: image

alexlioralexli commented 4 years ago

Downgrading to mujoco-py==1.50.1.68 solved the visualization problem for me. Note that Gym environments usually expect MuJoCo 150, as discussed here: https://github.com/openai/gym/issues/1541

jkterry1 commented 2 years ago

PR #2762 is about to be merged, introducing V4 MuJoCo environments using new bindings and a dramatically newer version of the engine. If this issue still persists with the V4 ones, please create a new issue for it.