Light Logo Dark Logo Minari Documentation
Farama Foundation logo Farama Foundation
Core Projects
  • Gymnasium logoGymnasium
  • PettingZoo logoPettingZoo
  • Minari logoMinari
Mature Projects
Documentation
  • Minigrid logoMinigrid
  • Gymnasium-Robotics logoGymnasium-Robotics
  • MAgent2 logoMAgent2
  • Miniworld logoMiniworld
  • MiniWoB++ logoMiniWoB++
  • Shimmy logoShimmy
  • MO-Gymnasium logoMO-Gymnasium
  • MOMAland logoMOMAland
Repositories
  • SuperSuit logoSuperSuit
  • Jumpy logoJumpy
Incubating Projects
Documentation
  • Arcade Learning Environment logoArcade Learning Environment
  • ViZDoom logoViZDoom
  • Metaworld logoMetaworld
  • HighwayEnv logoHighwayEnv
  • A2Perf logoA2Perf
  • MPE2 logoMPE2
Repositories
  • Procgen2 logoProcgen2
  • CrowdPlay logoCrowdPlay
  • Stable-Retro logoStable-Retro
  • ChatArena logoChatArena
Foundation
  • About logoAbout
  • Standards logoStandards
  • Donate logoDonate
Contents Menu Expand Light mode Dark mode Auto light/dark mode
Light Logo Dark Logo Minari Documentation

Introduction

  • Basic Usage
  • Minari CLI
  • Dataset Standards

API

  • Minari
    • MinariDataset
    • MinariStorage
    • EpisodeData
    • StepData
    • Namespace
  • DataCollector
    • EpisodeBuffer
    • StepDataCallback
    • EpisodeMetadataCallback

Tutorials

  • Dataset Creation
    • Collecting a subset of a dictionary space with StepDataCallback
    • Serializing a custom space
    • PointMaze D4RL dataset
  • Using Datasets
    • Behavioral cloning with PyTorch
    • Implicit Q-Learning with TorchRL

Datasets

  • D4RL
    • Ant Maze
      • Medium-Play
      • Umaze-Diverse
      • Large-Diverse
      • Large-Play
      • Medium-Diverse
      • Umaze
    • Point Maze
      • Open-Dense
      • Umaze
      • Large-Dense
      • Medium
      • Umaze-Dense
      • Medium-Dense
      • Large
      • Open
    • Pen
      • Human
      • Cloned
      • Expert
    • Kitchen
      • Partial
      • Complete
      • Mixed
    • Door
      • Human
      • Cloned
      • Expert
    • Hammer
      • Cloned
      • Expert
      • Human
    • Relocate
      • Human
      • Expert
      • Cloned
    • MiniGrid
      • Fourrooms-Random
      • Fourrooms
  • Minigrid
    • BabyAI Pickup
      • Optimal
      • Optimal-Fullobs
    • BabyAI PickupAbove
      • Optimal-Fullobs
      • Optimal
    • BabyAI PickupLoc
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToObjMazeOpen
      • Optimal
      • Optimal-Fullobs
    • BabyAI PutNextS7N4
      • Optimal
      • Optimal-Fullobs
    • BabyAI MoveTwoAcrossS5N2
      • Optimal-Fullobs
      • Optimal
    • BabyAI OpenDoor
      • Optimal
      • Optimal-Fullobs
    • BabyAI KeyCorridorS4R3
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToObjMazeS4R2
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToLocalS6N3
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToRedBallNoDists
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToLocalS8N2
      • Optimal
      • Optimal-Fullobs
    • BabyAI UnlockLocal
      • Optimal
      • Optimal-Fullobs
    • BabyAI KeyCorridorS3R2
      • Optimal-Fullobs
      • Optimal
    • BabyAI KeyCorridorS3R1
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToObjS4
      • Optimal
      • Optimal-Fullobs
    • BabyAI KeyCorridorS6R3
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToRedBall
      • Optimal-Fullobs
      • Optimal
    • BabyAI OpenTwoDoors
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToObjMazeS5
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToSeqS5R2
      • Optimal
      • Optimal-Fullobs
    • BabyAI SynthLoc
      • Optimal-Fullobs
      • Optimal
    • BabyAI KeyCorridorS5R3
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToObjMazeS7
      • Optimal-Fullobs
      • Optimal
    • BabyAI BossLevelNoUnlock
      • Optimal-Fullobs
      • Optimal
    • BabyAI KeyCorridorS3R3
      • Optimal-Fullobs
      • Optimal
    • BabyAI SynthSeq
      • Optimal
      • Optimal-Fullobs
    • BabyAI Synth
      • Optimal
      • Optimal-Fullobs
    • BabyAI ActionObjDoor
      • Optimal-Fullobs
      • Optimal
    • BabyAI UnlockPickup
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToLocalS5N2
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToLocal
      • Optimal
      • Optimal-Fullobs
    • BabyAI UnlockLocalDist
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToLocalS6N4
      • Optimal-Fullobs
      • Optimal
    • BabyAI BlockedUnlockPickup
      • Optimal-Fullobs
      • Optimal
    • BabyAI OneRoomS12
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToRedBlueBall
      • Optimal
      • Optimal-Fullobs
    • BabyAI OpenDoorDebug
      • Optimal-Fullobs
      • Optimal
    • BabyAI FindObjS7
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToSeq
      • Optimal
      • Optimal-Fullobs
    • BabyAI UnblockPickup
      • Optimal-Fullobs
      • Optimal
    • BabyAI OpenDoorsOrderN2Debug
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToObj
      • Optimal-Fullobs
      • Optimal
    • BabyAI PutNextS5N1
      • Optimal
      • Optimal-Fullobs
    • BabyAI FindObjS5
      • Optimal-Fullobs
      • Optimal
    • BabyAI KeyCorridor
      • Optimal
      • Optimal-Fullobs
    • BabyAI OpenDoorColor
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToLocalS8N4
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToOpen
      • Optimal
      • Optimal-Fullobs
    • BabyAI OpenDoorsOrderN4
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToLocalS8N7
      • Optimal-Fullobs
      • Optimal
    • BabyAI OpenDoorLoc
      • Optimal
      • Optimal-Fullobs
    • BabyAI PutNextS4N1
      • Optimal
      • Optimal-Fullobs
    • BabyAI FindObjS6
      • Optimal
      • Optimal-Fullobs
    • BabyAI OpenRedBlueDoors
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToLocalS6N2
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoTo
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToLocalS8N3
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToLocalS8N6
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToObjMazeS4
      • Optimal
      • Optimal-Fullobs
    • BabyAI Open
      • Optimal-Fullobs
      • Optimal
    • BabyAI PutNextS5N2
      • Optimal-Fullobs
      • Optimal
    • BabyAI OpenRedDoor
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToRedBallGrey
      • Optimal-Fullobs
      • Optimal
    • BabyAI MoveTwoAcrossS8N9
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToObjS6
      • Optimal
      • Optimal-Fullobs
    • BabyAI OneRoomS8
      • Optimal
      • Optimal-Fullobs
    • BabyAI BossLevel
      • Optimal-Fullobs
      • Optimal
    • BabyAI PutNextLocal
      • Optimal
      • Optimal-Fullobs
    • BabyAI OpenDoorsOrderN2
      • Optimal
      • Optimal-Fullobs
    • BabyAI PutNextS6N3
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToObjMaze
      • Optimal
      • Optimal-Fullobs
    • BabyAI OneRoomS20
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToLocalS8N5
      • Optimal-Fullobs
      • Optimal
    • BabyAI PutNextLocalS5N3
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToLocalS7N5
      • Optimal-Fullobs
      • Optimal
    • BabyAI UnlockPickupDist
      • Optimal
      • Optimal-Fullobs
    • BabyAI OneRoomS16
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToObjMazeS6
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToObjDoor
      • Optimal-Fullobs
      • Optimal
    • BabyAI OpenRedBlueDoorsDebug
      • Optimal
      • Optimal-Fullobs
    • BabyAI OpenDoorsOrderN4Debug
      • Optimal-Fullobs
      • Optimal
    • BabyAI PutNextLocalS6N4
      • Optimal
      • Optimal-Fullobs
    • BabyAI GoToLocalS7N4
      • Optimal-Fullobs
      • Optimal
    • BabyAI GoToDoor
      • Optimal
      • Optimal-Fullobs
  • MuJoCo
    • Ant
      • Expert

Development

  • Github
  • Release Notes
  • Contribute to the Docs
Back to top
Edit this page

Using Datasets¶

Behavioral cloning with PyTorch

Behavioral cloning with PyTorch

Behavioral cloning with PyTorch
Implicit Q-Learning with TorchRL

Implicit Q-Learning with TorchRL

Implicit Q-Learning with TorchRL
Next
Behavioral cloning with PyTorch
Previous
PointMaze D4RL dataset
Copyright © 2022

This page uses Google Analytics to collect statistics.

Versions