app.py

from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
import gradio as gr

model_path = "./tinyllama-qlora-support-bot-faq"

tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForCausalLM.from_pretrained(model_path)
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

def generate_response(instruction):
    prompt = f"### Instruction:\n{instruction}\n\n### Response:\n"
    output = pipe(prompt, max_new_tokens=100, do_sample=True, temperature=0.7)
    return output[0]['generated_text'].replace(prompt, "").strip()

gr.Interface(
    fn=generate_response,
    inputs=gr.Textbox(lines=3, placeholder="Ask your customer support question here..."),
    outputs=gr.Textbox(lines=6),
    title="🛠️ Customer Support Chatbot (TinyLlama + QLoRA)",
    description="Ask any support question. Model trained on MakTek/Customer_support_faqs_dataset using TinyLlama 1.1B."
).launch()