README.md

metadata

tags:
  - FrozenLake-v1-4x4-slippery
  - q-learning
  - reinforcement-learning
  - custom-implementation
model-index:
  - name: q-table-frozen-lake
    results:
      - task:
          type: reinforcement-learning
          name: reinforcement-learning
        dataset:
          name: FrozenLake-v1-4x4-slippery
          type: FrozenLake-v1-4x4-slippery
        metrics:
          - type: mean_reward
            value: 0.75 +/- 0.43
            name: mean_reward
            verified: false

Q-Learning Agent playing FrozenLake-v1

This is a trained Q-Learning agent playing FrozenLake-v1.

Usage

import gymnasium as gym
from huggingface_hub import snapshot_download

# https://github.com/libertininick/r2seedo
from r2seedo.io import load_n_verify_model

# Download model snapshot from Hugging Face Hub
repo_local_path = snapshot_download(
  repo_id="libertininick/q-table-frozen-lake", 
  local_dir="path/to/download",
)

# Load the model from the snapshot
model = load_n_verify_model(repo_local_path)

# Create the environment
env = env_slippery = gym.make(
  id='FrozenLake-v1',
  map_name='4x4',
  is_slippery=True,
)