Major Reinforcement Algorithms are implemented and benchmarked. Code is modular and simplistic, can be easily tinkered around to change characteristics of environment, model and hyperparameters.
Following algorithms are implemented:
- Deep-Q network (Implemented)
- Reinforce (Implemented)
- Deep deterministic policy gradient (In progress)
- Actor-Critic (In progress)
A detailed description of algorithm can be found in respective directories
Will be updated soon.