A set of simple problems solved using (D)RL
- Nothing so far
Feature: Implement a customDQN
to solve theMountainCar_v0
environment- Feature: Implement a custom
policy-based
model to solve theCartPole_v1
environment - Feature: Implement a custom
A2C
model to solve theCartPole_v1
environment Feature: Implement a vision-basedDQN
to solve theMountainCar_v0
enviroment- Feature: Implement a vision-based
A2C
to solve theCartPole_v1
enviroment Feature: SolveMoon Lander
using a custom agent- Feature: Add model architecture visualization to each project
- Feature: Add train and test accuracy plots for each project
- Feature: Add performance benchmark to all projects
- Feature: Solve
MountainCar-v0
without reward-shaping - Feature: Implement
Prioritized Experience Replay (PER)
- Feature: Solve some of the
MuJoCo
environments
- Hamid Mohammadi: [email protected]