Mostafa134/sac-LunarLander-v3

sac_LunarLander-v3_best

This is a custom SAC (Soft Actor-Critic) model trained on LunarLander-v3.

Model Details

• Architecture: SAC
• Environment: LunarLander-v3
• State Dimension: 8
• Action Dimension: 2

Usage

import torch
import gymnasium as gym
from SAC import SACAgent

# Load the model
env = gym.make("LunarLander-v3", continuous=True)
state_dim = env.observation_space.shape[0]
action_dim = env.action_space.shape[0]
config = {'learning_rate': 0.0003, 'gamma': 0.99, 'tau': 0.005, 'alpha': 0.2, 'batch_size': 256, 'buffer_size': 100000}

agent = SACAgent(state_dim, action_dim, env.action_space, config)
checkpoint = torch.load("model.pth", map_location="cpu")
agent.policy.load_state_dict(checkpoint['policy'])
if 'critic' in checkpoint:
    agent.critic.load_state_dict(checkpoint['critic'])
if 'critic_target' in checkpoint:
    agent.critic_target.load_state_dict(checkpoint['critic_target'])