app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
import spaces
# Define the model path
model_path = "mrcuddle/Ministral-Instruct-2410-8B-DPO-RP"

# Load the tokenizer and model
tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForCausalLM.from_pretrained(
    model_path,
    device_map="auto",
    torch_dtype='auto'
).eval()

@spaces.GPU
# Define the chatbot function
def chatbot(message, history):
    # Prepare the messages for the model
    messages = [
        {"role": "user", "content": message}
    ]

    # Tokenize the input
    input_ids = tokenizer.apply_chat_template(conversation=messages, tokenize=True, add_generation_prompt=True, return_tensors='pt')

    # Generate the response
    output_ids = model.generate(input_ids.to('cuda'))
    response = tokenizer.decode(output_ids[0][input_ids.shape[1]:], skip_special_tokens=True)

    # Append the response to the history
    history.append((message, response))
    return history, ""
# Create the Gradio interface
with gr.Blocks() as demo:
    gr.Markdown("## Chatbot using Transformers")

    # Create a Chatbot component
    chatbot_component = gr.Chatbot([], elem_id="chatbot")

    with gr.Row():
        # Create a Textbox component for user input
        txt = gr.Textbox(
            show_label=False,
            placeholder="Type your message here...",
        )

    # Attach the chatbot function to the Textbox component
    txt.submit(chatbot, [txt, chatbot_component], [chatbot_component, txt])

# Launch the Gradio interface
demo.launch()