Embedded Files
Agents are trained with 3 objects and evaluated on increasing numbers of objects.
PPP-Cube(Image)
PPP-Cube(Image)
Current Observation (Bottom) | Generated Subgoal (Middle) | Goal (Top)
Front View (Left) | Side View (Right)
4-Cubes
4-Cubes
5-Cubes
5-Cubes
Push-Tetris(Image)
Push-Tetris(Image)
Current Observation (Bottom) | Generated Subgoal (Middle) | Goal (Top)
4-Objects
4-Objects
5-Objects
5-Objects
PPP-Cube(State)
PPP-Cube(State)
Current Observation (Bottom) | Goal (Top)
6-Cubes
6-Cubes
Stack-Cube(State)
Stack-Cube(State)
Current Observation (Bottom) | Goal (Top)
4-Cubes-Stack-2
4-Cubes-Stack-2
Page updated
Google Sites
Report abuse