Upload folder using huggingface_hub

README.md

@@ -1,102 +1,10 @@
----
-tags:
-  - reinforcement-learning
-  - q-learning
-  - frozenlake
-license: mit
-library: gym
----
 
-# Q-Learning Model for FrozenLake
+    # **Q-Learning** Agent playing FrozenLake-v1-4x4-no_slippery
+    This is a trained model of a **Q-Learning** agent playing **FrozenLake-v1-4x4-no_slippery**.
 
-This model is a **Q-learning** agent trained to solve the **FrozenLake-v1** environment from OpenAI Gym.
-
-## Model Description
-
-The model uses Q-learning, a reinforcement learning algorithm, to navigate the FrozenLake environment. The agent learns by interacting with the environment, receiving rewards or penalties, and updating its Q-table accordingly.
-
-- **Environment**: FrozenLake-v1 (4x4 grid, no slippery surface)
-- **Algorithm**: Q-learning
-- **Action space**: 4 discrete actions (left, down, right, up)
-- **State space**: 16 discrete states (grid cells)
-- **Training duration**: Approximately [X hours] of training time.
-
-## Usage
-
-To use this model, you can load the trained Q-learning model from Hugging Face and run it in your environment.
-
-```python
-import gym
-from huggingface_hub import hf_hub_download
-import pickle
-
-# Load the model
-model_path = hf_hub_download(repo_id="willco-afk/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl")
-
-with open(model_path, 'rb') as f:
-    model = pickle.load(f)
-
-# Setup the environment
-env = gym.make("FrozenLake-v1", is_slippery=False)
-
-# Run your agent
-state = env.reset()
-done = False
-
-while not done:
-    action = model["qtable"].argmax(axis=1)[state]  # Choose the action with the highest Q-value
-    state, reward, done, info = env.step(action)
-
-    if done:
-        print(f"Episode finished with reward: {reward}")
-
-# Q-Learning Model for FrozenLake
-
-This model is a **Q-learning** agent trained to solve the **FrozenLake-v1** environment from OpenAI Gym.
-
-## Model Description
-
-The model uses Q-learning, a reinforcement learning algorithm, to navigate the FrozenLake environment. The agent learns by interacting with the environment, receiving rewards or penalties, and updating its Q-table accordingly.
-
-- **Environment**: FrozenLake-v1 (4x4 grid, no slippery surface)
-- **Algorithm**: Q-learning
-- **Action space**: 4 discrete actions (left, down, right, up)
-- **State space**: 16 discrete states (grid cells)
-- **Training duration**: Approximately [X hours] of training time.
-
-## Usage
-
-To use this model, you can load the trained Q-learning model from Hugging Face and run it in your environment.
-
-```python
-import gym
-from huggingface_hub import hf_hub_download
-import pickle
-
-# Load the model
-model_path = hf_hub_download(repo_id="willco-afk/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl")
-
-with open(model_path, 'rb') as f:
-    model = pickle.load(f)
-
-# Setup the environment
-env = gym.make("FrozenLake-v1", is_slippery=False)
-
-# Run your agent
-state = env.reset()
-done = False
-
-while not done:
-    action = model["qtable"].argmax(axis=1)[state]  # Choose the action with the highest Q-value
-    state, reward, done, info = env.step(action)
-
-    if done:
-        print(f"Episode finished with reward: {reward}")
-
-
-@misc{q-learning-frozenlake,
-  author = {William Copper},
-  title = {Q-Learning for FrozenLake-v1},
-  year = {2024},
-  howpublished = {\url{https://huggingface.co/willco-afk/q-FrozenLake-v1-4x4-noSlippery}},
-}
+    ## Usage
+    ```python
+    model = load_from_hub(repo_id="willco-afk/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl")
+    env = gym.make(model["env_id"])
+    ```
+    

q-learning.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:abba5d1a07c2d7ba7424095f86138828f035d28a27d86d5332135d8addc35ce2
-size 885
+oid sha256:9e98a2445193951bbe55c4d271cd3af3deb7f204d9f8899f7872794548ec2640
+size 914

results.json

@@ -0,0 +1 @@
+{"env_id": "FrozenLake-v1", "mean_reward": 1.0, "n_eval_episodes": 100, "eval_datetime": "2024-12-22T15:27:53.029538"}