app.py

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

# Check if CUDA is available
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
if device.type == "cpu":
    print("Warning: CUDA is not available. Running on CPU, which may be slow.")

# Load the tokenizer and model directly
model_name = "ruslanmv/ai-medical-model-32bit"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = model = AutoModelForCausalLM.from_pretrained(
    model_name, 
    device_map="auto", 
    load_in_8bit=True
).to(device)



# Function to ask medical questions
def ask_medical_question(question):
    prompt = f"<|start_header_id|>system<|end_header_id|> You are a Medical AI chatbot assistant. <|eot_id|><|start_header_id|>User: <|end_header_id|>This is the question: {question}<|eot_id|>"

    inputs = tokenizer(prompt, return_tensors="pt").to(device)
    outputs = model.generate(
        **inputs,
        max_new_tokens=100,
        temperature=0.7,
        do_sample=True,
        top_p=0.9,
        top_k=30,
    )
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return response


# Set up Gradio interface
iface = gr.Interface(fn=ask_medical_question, inputs="text", outputs="text")
iface.launch()