AuroraMiddleware/bullet3
1
0
Fork 0
mirror of https://github.com/bulletphysics/bullet3 synced 2024-12-13 05:10:08 +00:00
Code Issues Releases Wiki Activity

Explicitly state the datatypes in creating action/obs spaces

Browse Source 
This should get rid of the annoying warnings that pop up whenever an environment is created. Numpy's default datatype is float64, gym's is float32, the env is actually float32, but the initialization doesn't make it explicit so it causes redundant warnings.
```
/path/lib/python3.9/site-packages/gym/logger.py:34: UserWarning: WARN: Box bound precision lowered by casting to float32 warnings.warn(colorize("%s: %s" % ("WARN", msg % args), "yellow"))
```
This commit is contained in:
Ariel Kwiatkowski 2021-09-29 23:10:18 +02:00 committed by GitHub
parent ce26271923
commit 496e614f77
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
examples/pybullet/gym/pybullet_envs/robot_bases.py
View File

@ -22,10 +22,10 @@ class XmlBasedRobot:
self.ordered_joints = None
self.robot_body = None
high = np.ones([action_dim])
self.action_space = gym.spaces.Box(-high, high)
high = np.inf * np.ones([obs_dim])
self.observation_space = gym.spaces.Box(-high, high)
high = np.ones([action_dim], dtype=np.float32)
self.action_space = gym.spaces.Box(-high, high, dtype=np.float32)
high = np.inf * np.ones([obs_dim], dtype=np.float32)
self.observation_space = gym.spaces.Box(-high, high, dtype=np.float32)
#self.model_xml = model_xml
self.robot_name = robot_name