refactor: rename app from "Valluvar or AI?" to "ValluvarAI" and reorganize UI

Rename project title from "Valluvar or AI?" to "ValluvarAI" across README, app.py, and game_engine.py. Restructure Gradio interface by moving model details into collapsible "Model Card" and "Training Details" accordions at the top. Consolidate "About" tab to focus on capabilities and acknowledgements. Add 6 new theme quick-select buttons (கல்வி, காதல், பொருள், அறம், வீரம், வாய

README.md

@@ -1,5 +1,5 @@
 ---
-title: Valluvar or AI?
+title: ValluvarAI
 emoji: 🕉️
 colorFrom: red
 colorTo: purple
@@ -10,7 +10,7 @@ app_file: app.py
 pinned: false
 ---
 
-# Valluvar or AI? 🕉️
+# ValluvarAI 🕉️
 
 An AI that writes new Thirukkurals in the style of Thiruvalluvar.
 

app.py

@@ -1,4 +1,4 @@
-"""Gradio UI for Thirukkural GPT - Valluvar or AI."""
+"""Gradio UI for ValluvarAI."""
 
 # Patch Jinja2 LRUCache to handle unhashable keys (gradio 4.44.0 bug).
 try:
@@ -48,13 +48,37 @@ from model_engine import model, stoi, itos
 from kural_engine import generate_kural
 from game_engine import new_round, check_guess
 
-with gr.Blocks(title="Valluvar or AI?") as demo:
-    gr.Markdown("# 🕉️ Valluvar or AI?")
+with gr.Blocks(title="ValluvarAI") as demo:
+    gr.Markdown("# 🕉️ ValluvarAI")
     gr.Markdown(
         "An AI that writes new Thirukkurals in the style of Thiruvalluvar. "
         "Enter a Tamil theme to generate bilingual wisdom."
     )
 
+    with gr.Accordion("Model Card", open=False):
+        gr.Markdown(
+            f"""
+            **Architecture:** GPT ({model.config.n_layer}L/{model.config.n_head}H/{model.config.n_embd}D)  
+            **Parameters:** {sum(p.numel() for p in model.parameters()) / 1e6:.1f}M  
+            **Vocabulary:** {len(stoi)} characters (Tamil + English)  
+            **Tokenization:** Character-level  
+            **Training Steps:** 10,000  
+            **Device:** Apple MPS (Mac Mini)  
+            **Training Time:** ~5 hours  
+            **Final Loss:** ~1.5
+            """
+        )
+
+    with gr.Accordion("Training Details", open=False):
+        gr.Markdown(
+            f"""
+            **Steps:** 10,000
+            **Device:** Apple MPS (Mac Mini)
+            **Time:** ~5 hours
+            **Final Loss:** ~1.5
+            """
+        )
+
     with gr.Tab("✨ Generate Kural"):
         with gr.Row():
             with gr.Column():
@@ -97,6 +121,13 @@ with gr.Blocks(title="Valluvar or AI?") as demo:
                 "நட்பு",
                 "அரசியல்",
                 "அறன் வலியுறுத்தல்",
+                "கல்வி",
+                "காதல்",
+                "பொருள்",
+                "அறம்",
+                "வீரம்",
+                "வாய்மை",
+                "அன்பு",
             ]
             for theme in themes:
                 btn = gr.Button(theme)
@@ -152,20 +183,7 @@ with gr.Blocks(title="Valluvar or AI?") as demo:
     with gr.Tab("📊 About"):
         gr.Markdown(
             f"""
-            ## Model Details
-
-            - **Architecture:** GPT ({model.config.n_layer}L/{model.config.n_head}H/{model.config.n_embd}D)
-            - **Parameters:** {sum(p.numel() for p in model.parameters()) / 1e6:.1f}M
-            - **Vocabulary:** {len(stoi)} characters (Tamil + English)
-            - **Training Data:** Thirukkural (1330 kurals with English translations)
-            - **Tokenization:** Character-level
-
-            ## Training
-
-            - Steps: 10,000
-            - Device: Apple MPS (Mac Mini)
-            - Time: ~5 hours
-            - Final Loss: ~1.5
+            ValluvarAI is a GPT model trained on the **only** Thirukkural dataset.
 
             ## Capabilities
 
@@ -173,6 +191,8 @@ with gr.Blocks(title="Valluvar or AI?") as demo:
             - ✅ Produce coherent English translations
             - ✅ Handle traditional themes (virtue, politics, love)
             - ❌ Modern topics (science, technology) - not in training data
+            - ❌ Contemporary language - only classical Tamil
+            - ⚠️ May contain inaccuracies or repetitions
 
             ## Examples of AI vs Original
 
@@ -183,9 +203,7 @@ with gr.Blocks(title="Valluvar or AI?") as demo:
 
             ## Acknowledgements
 
-            - Tamil text and English translations courtesy of
-            [Project Madurai](https://www.projectmadurai.org/pm_etexts/utf8/pmuni0153.html)
-            - Special thanks to Project Madurai for making the Thirukkural text and translations available.
+            - Special thanks to [Project Madurai](https://www.projectmadurai.org/pm_etexts/utf8/pmuni0153.html) for making the Thirukkural text and translations available.
             """
         )
 

game_engine.py

@@ -1,4 +1,4 @@
-"""Game logic for Valluvar or AI guessing game."""
+"""Game logic for ValluvarAI guessing game."""
 import random
 from typing import Any, Final
 

hf-space/README.md

@@ -1,5 +1,5 @@
 ---
-title: Valluvar or AI?
+title: ValluvarAI
 emoji: 🕉️
 colorFrom: orange
 colorTo: red
@@ -10,7 +10,7 @@ app_file: app.py
 pinned: false
 ---
 
-# Valluvar or AI? 🕉️
+# ValluvarAI 🕉️
 
 An AI that writes new Thirukkurals in the style of Thiruvalluvar.
 

hf-space/game_engine.py

@@ -1,4 +1,4 @@
-"""Game logic for Valluvar or AI guessing game."""
+"""Game logic for ValluvarAI guessing game."""
 import random
 from typing import Any, Final