Conversation
emailweixu
left a comment
emailweixu
left a comment
There was a problem hiding this comment.
Just some general comments. Not yet going to implementation detail yet.
| max_episode_steps=1000, | ||
| gym_env_wrappers=(), | ||
| alf_env_wrappers=[AlfEnvironmentDMC2GYMWrapper], | ||
| image_channel_first=False): |
There was a problem hiding this comment.
since the observation of dmc is known to be channel first. We don't need this option.
| # See the License for the specific language governing permissions and | ||
| # limitations under the License. | ||
|
|
||
| import alf |
There was a problem hiding this comment.
Please add unittest suite_dmc_test to test its functionality and add it to .ci-cd/build.sh
| @@ -0,0 +1,245 @@ | |||
| # Copyright (c) 2021 Horizon Robotics and ALF Contributors. All Rights Reserved. | |||
There was a problem hiding this comment.
It's better to rename this file as suite_dmc.py
There was a problem hiding this comment.
This is independent of CURL. So it should be submitted in a separate PR.
|
|
||
|
|
||
|
|
| For installation of MuJoCo200, see https://roboti.us. | ||
| Args: | ||
| environment_name (str): Do not use this arg, this arg is here to | ||
| metch up with create_environment. |
| max_episode_steps (int): If None the max_episode_steps will be set to the | ||
| default step limit defined in the environment's spec. No limit is applied | ||
| if set to 0 or if there is no max_episode_steps set in the environment's | ||
| spec. |
There was a problem hiding this comment.
This comment needs to be updated.
|
|
||
|
|
||
| @alf.configurable | ||
| def wrap_env(gym_env, |
There was a problem hiding this comment.
It looks exact same as the suite_gym.wrap_env. Should use that instead.
| the end of encoder. | ||
|
|
||
| Retrun: | ||
| A CURL model. |
There was a problem hiding this comment.
no need to comment return for __init__. It always return an instance of the class.
| self._encoding_net = creat_encoder(self.after_crop_spec, feature_dim) | ||
| self._target_encoding_net = self._encoding_net.copy( | ||
| name='target_encoding_net_ctor') | ||
| self.CrossEntropyLoss = nn.CrossEntropyLoss(reduction='none') |
There was a problem hiding this comment.
class member name is lower case words prefixed by _
2 participants
No description provided.