URL: price: 3RMB/hour GPU: GeForce RTX 2080 Ti [Read More]
Tags: cs


Mujoco is a powerful simulation engine and it is used in almost all reinforcement learning tasks. However, the documentation of Mujoco is quite lacking which may result in difficulties when we try to interact with this environment. In the following sections, I try to set out the observation and action space of some well known Mujoco environments. [Read More]
Tags: rl

Temporal Difference

```python class QLearningAgent(agent.BaseAgent): def agent_init(self, agent_init_info): self.num_actions = agent_init_info[“num_actions”] self.num_states = agent_init_info[“num_states”] self.epsilon = agent_init_info[“epsilon”] self.step_size = agent_init_info[“step_size”] = agent_init_info[“discount”] self.rand_generator = np.random.RandomState(agent_info[“seed”]) self.q = np.zeros((self.num_states, self.num_actions)) # The array of action-value estimates. [Read More]
Tags: rl

Q Learning and SARSA

%matplotlib inline import numpy as np from scipy.stats import sem import matplotlib.pyplot as plt from rl_glue import RLGlue import agent import cliffworld_env from tqdm import tqdm import pickle [Read More]