Skip to content
Snippets Groups Projects
Select Git revision
  • 42-run-baselines-in-ci
  • 62-increase-unit-test-coverage
  • master default
  • patch-1
  • spm/pre-release-refactor
  • v0.2.0
  • v0.1.2
  • QT_Rendering
  • pre_agents_handles
  • v0.1.1
10 results
You can move around the graph by using the arrow keys.
Created with Raphaël 2.2.017May161514131211109876543130Apr29282726252423222120191817161512111098543first implementation of local observation, need testingbutchered Player to remove all trace of nn agent and torchadded test file - forgotten last timefixed Player in play_model to run with simple obs, and use random actionsBUG Fix: Fixed missing branches in Tree Observation visualizationminor changes to rendering of observationfixed flake8 errorsfixed formattingfixed bugs in get() funciton in observation_builderfixed bug in rendering of tree state observationupdated test_environments.py to import correct libraryadded initial tree renderingMerge branch 'master' of gitlab.aicrowd.com:flatland/flatlandupdated test for observation to account for new structure in repoMerge branch 'master' of gitlab.aicrowd.com:flatland/flatlandfixing msgpack serialization probs with numpy arrays appearing in EnvAgentMerge branch 'master' of gitlab.aicrowd.com:flatland/flatlandAdded visited list to observations.py Tree observation#26 pydepsMerge branch 'master' of gitlab.aicrowd.com:flatland/flatlandadding dt2.py - illustration of QT SVG render prob on windows#26 pydepsMerge branch 'master' of gitlab.aicrowd.com:flatland/flatlandnew training results#26 pydepsrefactored env observations ISSUE #24 + fixed most pylint errorsrefactored env observations ISSUE #24 + fixed most pylint errorsfixing flake8 errormerge commitusing msgpack to save/load stateUpdated training: New state has two time framesupdate requirements_dev.txtadded basic colors (from MPL colormap hsv) to SVGSVG trains bend around cornersadded SVG artwork in QTSVG renderer.Revert "trying to fix simulation infeaasibility"trying to fix simulation infeaasibilitylatest training resultsfixed missing loop over agent when episode is solvedfixed formatting