danijar/director
Deep Hierarchical Planning from Pixels
This project helps AI researchers and engineers train reinforcement learning agents to solve complex, long-horizon tasks more efficiently. It takes raw visual observations (pixels) as input and outputs a trained agent capable of breaking down tasks into subgoals and executing low-level actions to achieve them. This is primarily for those developing and evaluating advanced AI decision-making systems.
117 stars. No commits in the last 6 months.
Use this if you are developing AI agents for tasks that require sequential decision-making over long periods, especially when rewards are sparse and direct supervision is difficult.
Not ideal if you are looking for a pre-trained model for immediate deployment or if your tasks are simple and short-term, requiring only basic reinforcement learning approaches.
Stars
117
Forks
27
Language
Python
License
—
Category
Last pushed
Dec 21, 2022
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/danijar/director"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
LucasAlegre/sumo-rl
Reinforcement Learning environments for Traffic Signal Control with SUMO. Compatible with...
hilo-mpc/hilo-mpc
HILO-MPC is a Python toolbox for easy, flexible and fast development of...
reiniscimurs/DRL-robot-navigation
Deep Reinforcement Learning for mobile robot navigation in ROS Gazebo simulator. Using Twin...
kyegomez/RoboCAT
Implementation of Deepmind's RoboCat: "Self-Improving Foundation Agent for Robotic Manipulation"...
OpenQuadruped/spot_mini_mini
Dynamics and Domain Randomized Gait Modulation with Bezier Curves for Sim-to-Real Legged Locomotion.