Update app.py

app.py

@@ -2,72 +2,61 @@ import gradio as gr
 from huggingface_hub import hf_hub_download
 from llama_cpp import Llama
 
-# 📥 1. Infinity Engine (GGUF Model) Download
-# RockSky1 ki repo se model file fetch ho rahi hai
+# 📥 Model Download
 model_path = hf_hub_download(
-    repo_id="RockSky1/Infinity_1.0", 
+    repo_id="RockSky1/Infinity_1.0",
     filename="Infinity_1.0.gguf"
 )
 
-# 🧠 2. Model Loading (Optimized for Free Tier)
-# n_ctx=512 aur n_threads=1 isliye rakha hai taaki OOM crash na ho
+print("Loading model...")
+
 llm = Llama(
-    model_path=model_path, 
-    n_ctx=512, 
-    n_threads=1
+    model_path=model_path,
+    n_ctx=512,
+    n_threads=2,
+    n_batch=128
 )
 
+print("Model loaded ✅")
+
+
 def chat_function(message, history):
-    # 🎭 System Prompt: Personal Branding & Identity
-    system_instruction = (
-        "You are Infinity 1.0, a powerful and futuristic AI engine. "
-        "You were developed by the expert AI Architect Shivam Kumar, "
-        "also known as RockSky1, hailing from Bihar, India. "
-        "You are highly intelligent, logical, and helpful. Always give credit to your creator "
-        "Shivam Kumar if someone asks who made you. Represent the innovation of Bihar!"
-    )
-    
-    # Prompt Formatting for the Model
-    full_prompt = f"System: {system_instruction}\nUser: {message}\nInfinity:"
-    
-    # AI Response Generation
-    response = llm(
-        full_prompt, 
-        max_tokens=256, 
-        stop=["User:", "System:", "\n"], 
-        echo=False
-    )
-    
-    return response["choices"][0]["text"].strip()
+    try:
+        output = llm.create_chat_completion(
+            messages=[
+                {
+                    "role": "system",
+                    "content": "You are Infinity AI 🔥 created by Shivam Kumar (RockSky1) from Bihar, India."
+                },
+                *[
+                    {"role": "user", "content": h[0]} if i % 2 == 0
+                    else {"role": "assistant", "content": h[1]}
+                    for i, h in enumerate(history)
+                ],
+                {"role": "user", "content": message}
+            ],
+            max_tokens=128,
+            temperature=0.7
+        )
+
+        reply = output["choices"][0]["message"]["content"]
+
+    except Exception as e:
+        reply = f"Error: {str(e)}"
+
+    return reply
+
+
+with gr.Blocks() as demo:
+    gr.Markdown("# ♾️ Infinity-LLM v1.0 🚀")
 
-# ✨ 3. Premium UI with Markdown & Emojis
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("""
-    # ♾️ Infinity-LLM v1.0
-    ### 🚀 Next-Gen Neural Engine by **Shivam Kumar (RockSky1)**
-    📍 *Patna, Bihar, India* 🇮🇳
-    
-    ---
-    **Welcome to the future.** Infinity-LLM is built for high-speed logic and creative intelligence. 
-    Ask me anything about technology, coding, or my creator!
-    """)
-    
     gr.ChatInterface(
         fn=chat_function,
         examples=[
-            "Who is Shivam Kumar?", 
-            "Tell me about Infinity-LLM.", 
+            "Who is Shivam Kumar?",
+            "Tell me about Infinity-LLM.",
             "Write a simple Python script."
-        ],
-        cache_examples=False,
+        ]
     )
-    
-    gr.Markdown("""
-    ---
-    *Built with ❤️ by RockSky1 | Powered by Infinity AI Ecosystem*
-    """)
 
-# 🏁 4. Launching the App
-if __name__ == "__main__":
-    demo.launch()
-    
+demo.launch()