Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,73 @@
+---
+tags:
+- SpaceInvadersNoFrameskip-v4
+- deep-reinforcement-learning
+- reinforcement-learning
+model-index:
+- name: PPO
+  results:
+  - task:
+      type: reinforcement-learning
+      name: reinforcement-learning
+    dataset:
+      name: SpaceInvadersNoFrameskip-v4
+      type: SpaceInvadersNoFrameskip-v4
+    metrics:
+    - type: mean_reward
+      value: 900.0
+      name: mean_reward
+      verified: false
+---
+
+# PPO Agent playing SpaceInvadersNoFrameskip-v4
+
+This is a trained model of a PPO agent playing SpaceInvadersNoFrameskip-v4 using CleanRL.
+
+## Metrics
+- **Mean Reward**: 900.0
+
+## Usage
+```python
+import torch
+import gymnasium as gym
+from PPO_atari import Agent
+
+env = gym.make("SpaceInvadersNoFrameskip-v4")
+
+# Load the model
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+env = gym.make("SpaceInvadersNoFrameskip-v4")
+agent = Agent(env).to(device)
+agent.load_state_dict(torch.load("model.pth", map_location=device))
+agent.eval()
+
+# Run evaluation
+obs, _ = env.reset()
+done = False
+while not done:
+    action, _, _, _ = agent.get_action_and_value(torch.tensor(obs).unsqueeze(0).to(device))
+    obs, reward, terminated, truncated, _ = env.step(action.cpu().numpy()[0])
+    done = terminated or truncated
+```
+
+## Training Details
+
+- **Algorithm**: Proximal Policy Optimization (PPO)
+- **Environment**: SpaceInvadersNoFrameskip-v4
+- **Total timesteps**: 10,000,000
+- **Framework**: CleanRL
+- **Number of parallel environments**: 8
+- **Learning rate**: 2.5e-4
+- **Evaluation episodes**: 100
+- **Mean reward**: 900.00
+
+## Hyperparameters
+
+- Learning rate: 2.5e-4
+- Gamma: 0.99
+- GAE Lambda: 0.95
+- Clip coefficient: 0.1
+- Value function coefficient: 0.5
+- Entropy coefficient: 0.01
+- Number of epochs: 4
+- Minibatches: 4