Updated README.md

README.md

@@ -1,32 +1,42 @@
-%%writefile README.md
-# PPO SnowballTarget Agent
+---
+tags:
+- reinforcement-learning
+- ml-agents
+- ppo
+- unity
+- SnowballTarget
+license: mit
+---
 
-This model was trained using Proximal Policy Optimization (PPO) with Unity ML-Agents as part of the Hugging Face Deep Reinforcement Learning Course.
+# PPO SnowballTarget
 
-## Environment
-- Unity ML-Agents
-- SnowballTarget environment
+This is a trained PPO agent playing SnowballTarget using Unity ML-Agents.
 
-## Training Details
-- Algorithm: PPO
-- Total training steps: 200,000
-- Final mean reward: ~23.2
+## Environment
+SnowballTarget
 
-## Results
-The agent learned to consistently hit targets in the SnowballTarget environment and achieved stable rewards during training.
+## Algorithm
+PPO (Proximal Policy Optimization)
 
-Final training logs:
-- Step 160000 → Mean Reward: 22.84
-- Step 170000 → Mean Reward: 22.85
-- Step 180000 → Mean Reward: 23.00
-- Step 190000 → Mean Reward: 23.46
-- Step 200000 → Mean Reward: 23.21
+## Training Results
 
-## Files
-- `SnowballTarget.onnx` — trained Unity ML-Agents policy network
+Final mean reward: ~23.2 after 200k training steps.
 
 ## Usage
-This model can be loaded into Unity ML-Agents for inference and evaluation.
 
-## Author
-Ryan Aparicio
+You can watch the agent play directly in your browser:
+
+1. Go to:
+https://huggingface.co/spaces/ThomasSimonini/ML-Agents-SnowballTarget
+
+2. In the model selector, enter:
+
+RyanAA/ppo-SnowballTarget
+
+3. Select `SnowballTarget.onnx`
+
+4. Click "Watch the agent play"
+
+## Files
+
+- `SnowballTarget.onnx` — trained policy network