Skip to content
Snippets Groups Projects

refactor tree data structure and add unit test for the normalization of the features

Merged refactor tree data structure and add unit test for the normalization of the features
3-refactor-data-structure-of-treeobsforrailenv into master
2 unresolved threads
Merged Christian Baumberger requested to merge 3-refactor-data-structure-of-treeobsforrailenv into master
2 unresolved threads

Closes #3 (closed)

Merge request reports

Loading
Loading

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
torch_training/multi_agent_training.py
172 172
173 173 # Build agent specific observations
174 174 for a in range(env.get_num_agents()):
175 agent_obs[a] = agent_obs[a] = normalize_observation(obs[a], observation_radius=10)
175 agent_obs[a] = agent_obs[a] = normalize_observation(obs[a], tree_depth, observation_radius=10)
  • Erik Nygren
    Erik Nygren @mlerik started a thread on the diff
    • utils/observation_utils.py
    53 54 return np.clip((np.array(obs) - min_obs) / norm, clip_min, clip_max)
    54 55
    55 56
    56 def split_tree(tree, num_features_per_node, current_depth=0):
    57 def _split_node_into_feature_groups(node: TreeObsForRailEnv.Node) -> (np.ndarray, np.ndarray, np.ndarray):
  • Erik Nygren merged

    merged

  • Erik Nygren mentioned in commit 4b9c3107

    mentioned in commit 4b9c3107

    • Please register or sign in to reply