app.py

import gradio as gr
from transformers import AutoModel, AutoTokenizer
import torch

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("mradermacher/Fimbulvetr-11B-v2-GGUF")
model = AutoModel.from_pretrained("mradermacher/Fimbulvetr-11B-v2-GGUF")

def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    messages = system_message + "\n" + "\n".join([f"User: {h[0]}\nBot: {h[1]}" for h in history if h]) + f"\nUser: {message}"
    
    inputs = tokenizer(messages, return_tensors="pt", truncation=True)
    with torch.no_grad():
        output = model.generate(
            **inputs,
            max_new_tokens=max_tokens,
            temperature=temperature,
            top_p=top_p,
            do_sample=True
        )
    response = tokenizer.decode(output[0], skip_special_tokens=True)
    yield response


demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a friendly storyteller.", label="System message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.95,
            step=0.05,
            label="Top-p (nucleus sampling)",
        ),
    ],
)

if __name__ == "__main__":
    demo.launch()