Mehooz / vision4leg

Toolkit for vision-guided quadrupedal locomotion research
222 stars 28 forks source link

Baseline environment returns wrong dimension #14

Open dibbla opened 1 year ago

dibbla commented 1 year ago

Hi! Thank you for open-sourcing the work.

I'm running the baseline script with python starter/ppo_nature_cnn_vision_only.py --config config/rl/static/naive_baseline/thin.json.

However, the environment keeps returning shape 65728 after setting add_last_action_input to False. This is incompatible with visual input shape [4, 4, 64, 64]. Do you know how to configure this?

Thanks!

Kartikgar commented 1 year ago

Hi, I am facing the same issue. Were you able to figure it out?