app.py

import gradio as gr

# Load your model and tokenizer
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Specify the model name
model_name = "ahmedbasemdev/llama-3.2-3b-ChatBot"

# Load the model with 8-bit quantization
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    device_map="auto",  # Automatically map the model to the available device (CPU)
    load_in_8bit=True,  # Enable 8-bit quantization
    torch_dtype=torch.float16  # Use mixed precision
)

# Load the tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_name)

def single_inference(question):
    messages = []

    messages.append({"role": "user", "content": question})

    input_ids = tokenizer.apply_chat_template(
        messages,
        add_generation_prompt=True,
        return_tensors="pt"
    ).to(model.device)

    terminators = [
        tokenizer.eos_token_id,
        tokenizer.convert_tokens_to_ids("<|eot_id|>")
    ]

    outputs = model.generate(
        input_ids,
        max_new_tokens=256,
        eos_token_id=terminators,
        do_sample=True,
        temperature=0.2,
    )
    response = outputs[0][input_ids.shape[-1]:]
    output = tokenizer.decode(response, skip_special_tokens=True)
    return output

# Create the Gradio interface
interface = gr.Interface(
    fn=single_inference,  # Function to wrap
    inputs=gr.Textbox(lines=2, placeholder="Ask a question..."),  # Input type
    outputs=gr.Textbox(label="Response"),  # Output type
    title="Chat with Your Model",  # App title
    description="Enter a question, and the model will generate a response.",  # App description
)

# Launch the app
if __name__ == "__main__":
    interface.launch()