03 / RL
RL
Reinforcement learning experiments
- 01 Snake RL
A snake agent trained from scratch with Dueling Double DQN. PyTorch, 28-dim egocentric state, target network, 256-unit network. Greedy eval mean: 102.
03 / RL
Reinforcement learning experiments
A snake agent trained from scratch with Dueling Double DQN. PyTorch, 28-dim egocentric state, target network, 256-unit network. Greedy eval mean: 102.