Deploy Llama model in AMD Radeon PRO W7900

import gradio
import requests


def run(prompt):
    url = "http://127.0.0.1:8099/completion"
    data = {"prompt": prompt, "n_predict": 128}
    resp = requests.post(url, json=data)
    data = resp.json()
    content = data.pop("content")
    return content, data


# demo = gradio.Interface(
#        fn=run,
#         inputs=["text"],
#        outputs=["text", "json"]
# )


with gradio.Blocks() as demo:
    inp = gradio.Textbox(
        label="Prompt",
        placeholder="prompt",
        lines=3
    )

    btn_run = gradio.Button("Run")
    out = gradio.Textbox()
    extra = gradio.Json()

    btn_run.click(fn=run, inputs=inp, outputs=[out, extra])

# demo.launch(server_name="0.0.0.0")
demo.launch(server_name="0.0.0.0", share=True)