app.py

import string
import gradio as gr
import requests
import io

url=""
def inference_chat(input_image,input_text):
    with io.BytesIO() as buf:
        input_image.save(buf, 'jpeg')
        image_bytes = buf.getvalue()
        files={"img":("input_imge.jpg",image_bytes,'image/png',{})}
        res=requests.request("post",url=url,data={"input_text":input_text},files=files)
        return res.json()["answer"]
    
with gr.Blocks(
    css="""
    .message.svelte-w6rprc.svelte-w6rprc.svelte-w6rprc {font-size: 20px; margin-top: 20px}
    #component-21 > div.wrap.svelte-w6rprc {height: 600px;}
    """
) as iface:
    state = gr.State([])
    #caption_output = None
    #gr.Markdown(title)
    #gr.Markdown(description)
    #gr.Markdown(article)

    with gr.Row():
        with gr.Column(scale=1):
            image_input = gr.Image(type="pil")
            with gr.Row():
                with gr.Column(scale=1):
                    chat_input = gr.Textbox(lines=1, label="VQA Input(问题输入)")
                    with gr.Row():
                        clear_button = gr.Button(value="Clear", interactive=True)
                        submit_button = gr.Button(
                            value="Submit", interactive=True, variant="primary"
                        )
        with gr.Column():
            caption_output = gr.Textbox(lines=0, label="VQA Output(模型答案输出)")
            
            
        image_input.change(
            lambda: ("", "", []),
            [],
            [ caption_output, state],
            queue=False,
        )
        chat_input.submit(
                    inference_chat,
                    [
                        image_input,
                        chat_input,
                    ],
                    [ caption_output],
                )
        clear_button.click(
                        lambda: ("", [], []),
                        [],
                        [chat_input,  state],
                        queue=False,
                    )
        submit_button.click(
                        inference_chat,
                        [
                            image_input,
                            chat_input,
                        ],
                        [caption_output],
                    )

   # examples = gr.Examples(
   #     examples=examples,
   #     inputs=[image_input, chat_input],
  #  )

iface.queue(concurrency_count=1, api_open=False, max_size=10)
iface.launch(enable_queue=True)