mirror of
https://github.com/gryf/coach.git
synced 2025-12-18 03:30:19 +01:00
875d6ef017
* Adding target reward * Adding target successs * Addressing comments * Using custom_reward_threshold and target_success_rate * Adding exit message * Moving success rate to environment * Making target_success_rate optional
A custom environment implementation should look like this:
from coach.filters.input_filter import InputFilter
class CustomFilter(InputFilter):
def __init__(self):
...
def _filter(self, env_response: EnvResponse) -> EnvResponse:
...
def _get_filtered_observation_space(self, input_observation_space: ObservationSpace) -> ObservationSpace:
...
def _get_filtered_reward_space(self, input_reward_space: RewardSpace) -> RewardSpace:
...
def _validate_input_observation_space(self, input_observation_space: ObservationSpace):
...
def _reset(self):
...