app.py

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

model_name = "redietmolla/amharic_qa_fine_tuned_llama_model"

# Load your fine-tuned model and tokenizer
model = AutoModelForCausalLM.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)

def qa_pipeline(question):
    qa_pipeline = pipeline("text-generation", model=model, tokenizer=tokenizer)
    result = qa_pipeline(f"<s>[INST] {question} [/INST]")
    return result[0]['generated_text']

iface = gr.Interface(
    fn=qa_pipeline,
    inputs=gr.Textbox(lines=2, placeholder="Enter your question here..."),
    outputs="text",
    title="Amharic QA Model",
    description="Ask questions and get answers based on the Amharic context."
)

iface.launch()