import os
os.environ["CUDA_VISIBLE_DEVICES"] = "-1"

import gradio as gr
from unsloth import FastLanguageModel
import torch

max_seq_length = 2048  # Choose any! We auto support RoPE Scaling internally!
dtype = None  # None for auto detection. Float16 for Tesla T4, V100, Bfloat16 for Ampere+
load_in_4bit = True  # Use 4bit quantization to reduce memory usage. Can be False.

# Force the model to run on CPU only by setting the device to "cpu"
device = "cpu"

# Load model and tokenizer with the device set to "cpu"
model, tokenizer = FastLanguageModel.from_pretrained(
    model_name="llama_lora_model_1",
    max_seq_length=max_seq_length,
    dtype=dtype,
    load_in_4bit=load_in_4bit,
)

# Move the model to CPU (even if it was initially loaded with GPU support)
model.to(device)

# Respond function
def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    # Prepare the system message
    messages = [{"role": "system", "content": system_message}]

    # Add history to the messages
    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})

    # Add the current message from the user
    messages.append({"role": "user", "content": message})

    # Prepare the inputs for the model
    inputs = tokenizer.apply_chat_template(
        messages,
        tokenize=True,
        add_generation_prompt=True,
        return_tensors="pt",
    )

    # Generate the response using your model on CPU
    outputs = model.generate(
        input_ids=inputs["input_ids"].to(device),  # Ensure input is on the CPU
        max_new_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p,
        use_cache=True,
    )

    # Decode the generated output
    response = tokenizer.batch_decode(outputs, skip_special_tokens=True)

    # Return the response
    return response[0]


# Gradio interface setup
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a friendly Chatbot.", 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()