Some questions about the paper #22
Description
In the low-level policy, the paper mentions that the observation is 90 dimensions, but in the open source code, the observation values seem to be inconsistent with the dimensions introduced in the paper. May I ask what the author's considerations are. Thank you!