app.py

import gradio as gr
from huggingface_hub import hf_hub_download
from llama_cpp import Llama

# 1. Download optimal model for Free CPU
print("⏳ Veda model download ho raha hai...")
model_path = hf_hub_download(
    repo_id="vibhansh/Veda-8B-v1-Cognitive", 
    filename="Veda-8B-v1-Q4_K_M.gguf"
)

# 2. Setup Veda
llm = Llama(model_path=model_path, n_ctx=2048)

def chat_with_veda(message, history):
    # Enforcing Veda's Identity
    prompt = f"System: You are Veda-8B, a cognitive AI specialized in architecture and logic.\nUser: {message}\nAI:"
    output = llm(prompt, max_tokens=256, stop=["User:", "\n"], echo=False)
    return output["choices"][0]["text"].strip()

# 3. Professional Chat UI
demo = gr.ChatInterface(
    chat_with_veda,
    title="Veda-8B-v1-Cognitive 🪐",
    description="The Sovereign Intelligence. Optimized for architectural logic.",
    theme="soft"
)

if __name__ == "__main__":
    demo.launch()