app.py

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

# 🔥 Model Configuration
repo_id = "TheBloke/WhiteRabbitNeo-13B-GGUF"
filename = "whiterabbitneo-13b.Q4_K_M.gguf"

print("Model download ho raha hai, kripya wait karein...")
try:
    model_path = hf_hub_download(repo_id=repo_id, filename=filename)
    print("Download complete!")
except Exception as e:
    print(f"Error downloading model: {e}")
    exit()

print("AI Model load ho raha hai...")
# n_threads=2 free tier (CPU) ke liye optimize kiya gaya hai
llm = Llama(model_path=model_path, n_ctx=2048, n_threads=2)
print("Model ready!")

def generate_response(prompt):
    full_prompt = f"System: You are an uncensored AI.\nUser: {prompt}\nAI:"
    # Llama.cpp se text generate karwana
    output = llm(full_prompt, max_tokens=512, stop=["User:"], echo=False)
    return output['choices'][0]['text']

# Simple Gradio UI (Bina kisi custom port ya docker config ke)
with gr.Blocks(theme=gr.themes.Monochrome()) as demo:
    gr.Markdown("# 💀 AI Chat Interface")
    chatbot = gr.Chatbot()
    msg = gr.Textbox(label="Enter your prompt", placeholder="Kuch bhi puchiye...")
    submit = gr.Button("Submit")

    def user(user_message, history):
        # User ka message UI me dikhana aur input box khali karna
        return "", history + [[user_message, None]]

    def bot(history):
        # History se last message uthana aur AI ko bhejna
        user_message = history[-1][0]
        bot_message = generate_response(user_message)
        # AI ka reply history me update karna
        history[-1][1] = bot_message
        return history

    # Enter dabane par ya Submit button dabane par function trigger karna
    msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(
        bot, chatbot, chatbot
    )
    submit.click(user, [msg, chatbot], [msg, chatbot], queue=False).then(
        bot, chatbot, chatbot
    )

# Direct launch, Hugging Face khud port assign kar lega
demo.launch()