app.py

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
import gradio as gr
import os

model_name = "unsloth/Phi-4-mini-reasoning-unsloth-bnb-4bit"

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)

model = AutoModelForCausalLM.from_pretrained(
    model_name,
    device_map="auto",
    torch_dtype=torch.float16,
    trust_remote_code=True
)

pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
)

def chat_fn(prompt: str) -> str:
    if not prompt.strip():
        return "Please enter a prompt."

    try:
        output = pipe(
            prompt,
            max_new_tokens=128,
            do_sample=False,
            pad_token_id=tokenizer.eos_token_id,
            return_full_text=False,
        )

        generated_text = output[0]["generated_text"]

        final_result = generated_text[:500].encode("utf-8", "ignore").decode("utf-8")

        if not final_result.strip():
            return "The model did not generate a response. Try a different prompt."

        return final_result

    except Exception as e:
        print(f"An error occurred during text generation: {e}")
        return f"[ERROR] An issue occurred while generating the response. Please try again or simplify your prompt. Details: {str(e)[:200]}"

demo_interface = gr.Interface(
    fn=chat_fn,
    inputs=gr.Textbox(lines=5, placeholder="Enter your prompt here...", label="Your Prompt"),
    outputs=gr.Textbox(lines=10, label="Generated Response"),
    title="Phi-4 Mini Reasoning Chatbot",
    description="Ask the Phi-4 Mini Reasoning model anything. Responses are limited to prevent errors.",
    allow_flagging="never",
)

demo_interface.launch()