app.py

import gradio as gr
from transformers import pipeline

pipe = pipeline("text-generation", model="ynanxiu/olmo3-190m-zh-nano", model_kwargs={"attn_implementation": "sdpa"})

def predict(message):
    output = pipe(message, max_new_tokens=256)
    return output[0]["generated_text"]

gr.Interface(fn=predict, inputs="text", outputs="text").launch()