app.py

import gradio as gr
import predictLLaVa

image = gr.Image(sources=["upload", "webcam"], label="Image", type="pil")
inputText = gr.Textbox(lines=1, label="English Instruction:")
languageDropdown = gr.Dropdown(
    choices=["en", "de", "fr", "es", "it", "nl", "ru"],
    label="Output in:",
    multiselect=False,
    value="en"
)

gr.Interface(
    fn=predictLLaVa.process_single_image,
    inputs=[inputText, image, languageDropdown],
    outputs=gr.Textbox(label="Generated Description"),
    title="Fluffyvision",
    description="Generate a text description of an image based on an instruction.",
    allow_flagging="never"
).launch(share=True, auth=predictLLaVa.check_auth, auth_message="Please enter your username and password.")