Update README.md

README.md

@@ -1,3 +1,43 @@
----
-license: mit
----
+---
+license: mit
+tags:
+  - sklearn
+  - tabular-regression
+  - distillation
+  - chemical-engineering
+---
+
+# Pyrolysis Distillation Predictor
+
+Predicts NAPTHA and DIESEL purity from distillation column operating conditions.
+
+## Inputs
+| Feature | Description |
+|---|---|
+| Distillate_To_Feed_Ratio | Ratio of distillate to feed flow |
+| Feed_Stage | Feed stage number |
+| top_stage_pressure_(bar) | Top stage pressure |
+| Temp_of_Field_(C) | Field temperature |
+| Feed_Flow_Rate_(Kg/hr) | Feed flow rate |
+
+## Outputs
+- `NAPTHA`: predicted purity (0–1)
+- `DIESEL`: predicted purity (0–1)
+
+## Feasible Operating Zone
+Both outputs ≥ 80% when Distillate_To_Feed_Ratio is between 0.20 and 0.44.
+
+## Usage
+```python
+import joblib
+from huggingface_hub import hf_hub_download
+
+path = hf_hub_download("your-username/pyrolysis-distillation-predictor", 
+                        "pyrolysis_model.joblib")
+model = joblib.load(path)
+
+prediction = model.predict([[0.35, 10, 2.5, 150, 1000]])
+print(f"NAPTHA: {prediction[0][0]:.3f}, DIESEL: {prediction[0][1]:.3f}")
+```
+
+One thing to keep in mind: HuggingFace doesn't natively "understand" sklearn models the way it understands transformers, so there's no automatic inference API. If you want a live demo, you'd pair it with a **Gradio Space** — which is actually very easy and looks great for a senior project presentation. Want me to write that too?