utils.py

import gradio as gr
import os
import torch
from transformers import AutoConfig, AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer
from threading import Thread
from accelerate import init_empty_weights, infer_auto_device_map, disk_offload

# Set environment variables
HF_TOKEN = os.getenv("HF_TOKEN")

DESCRIPTION = '''
<div>
<h1 style="text-align: center;">ContenteaseAI custom trained model</h1>
</div>
'''

LICENSE = """
<p/>

---
For more information, visit our [website](https://contentease.ai).
"""

PLACEHOLDER = """
<div style="padding: 30px; text-align: center; display: flex; flex-direction: column; align-items: center;">
   <h1 style="font-size: 28px; margin-bottom: 2px; opacity: 0.55;">ContenteaseAI Custom AI trained model</h1>
   <p style="font-size: 18px; margin-bottom: 2px; opacity: 0.65;">Enter the text extracted from the PDF:</p>
</div>
"""

css = """
h1 {
  text-align: center;
  display: block;
}
"""

def initialize_model(model_name, max_memory=None):
    device = torch.device('cpu')
    
    # Load model configuration
    config = AutoConfig.from_pretrained(model_name)
    
    with init_empty_weights():
        # Initialize model with empty weights
        model = AutoModelForCausalLM.from_config(config)
    
    # Create device map based on memory constraints
    device_map = infer_auto_device_map(
        model, max_memory=max_memory, no_split_module_classes=["GPTNeoXLayer"], dtype="float16"
    )
    
    # Determine if offloading is needed
    needs_offloading = any(device == 'disk' for device in device_map.values())
    
    if needs_offloading:
        # Load model for offloading
        model = AutoModelForCausalLM.from_pretrained(
            model_name, device_map=device_map, offload_folder="offload",
            offload_state_dict=True, torch_dtype=torch.float16
        )
        offload_directory = "offload/"
        # Offload model to disk
        disk_offload(model=model, offload_dir=offload_directory)
    else:
        # Load model normally to specified device
        model = AutoModelForCausalLM.from_pretrained(
            model_name, torch_dtype=torch.float16
        )
        model.to(device)
    
    return model

try:
    # Initialize the model and tokenizer
    model_name = "meta-llama/Meta-Llama-3-8B-Instruct"
    model = initialize_model(model_name, max_memory={"cpu": "GiB"})
    tokenizer = AutoTokenizer.from_pretrained(model_name, use_auth_token=HF_TOKEN)
except Exception as e:
    print(f"Error initializing model: {e}")
    exit(1)

terminators = [
    tokenizer.eos_token_id,
    tokenizer.convert_tokens_to_ids("")
]

def chat_llama3_8b(message: str, history: list, temperature: float, max_new_tokens: int) -> str:
    """
    Generate a streaming response using the llama3-8b model.
    Args:
        message (str): The input message.
        history (list): The conversation history used by ChatInterface.
        temperature (float): The temperature for generating the response.
        max_new_tokens (int): The maximum number of new tokens to generate.
    Returns:
        str: The generated response.
    """
    conversation = []
    message += " Extract all relevant keywords and add quantity from the following text and format the result in nested JSON:"
    for user, assistant in history:
        conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
    conversation.append({"role": "user", "content": message})

    input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt").to(model.device)
    
    streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)

    generate_kwargs = dict(
        input_ids=input_ids,
        streamer=streamer,
        max_new_tokens=max_new_tokens,
        do_sample=True,
        temperature=temperature,
        eos_token_id=terminators,
    )
    if temperature == 0:
        generate_kwargs['do_sample'] = False
        
    t = Thread(target=model.generate, kwargs=generate_kwargs)
    t.start()

    outputs = []
    for text in streamer:
        outputs.append(text)
        yield "".join(outputs)

# Gradio block
chatbot = gr.Chatbot(height=450, placeholder=PLACEHOLDER, label='Gradio ChatInterface')

with gr.Blocks(fill_height=True, css=css) as demo:
    gr.Markdown(DESCRIPTION)
    gr.ChatInterface(
        fn=chat_llama3_8b,
        chatbot=chatbot,
        fill_height=True,
        additional_inputs_accordion=gr.Accordion(label="⚙️ Parameters", open=False, render=False),
        additional_inputs=[
            gr.Slider(
                minimum=0,
                maximum=1,
                step=0.1,
                value=0.95,
                label="Temperature",
                render=False
            ),
            gr.Slider(
                minimum=128,
                maximum=9012,
                step=1,
                value=512,
                label="Max new tokens",
                render=False
            ),
        ]
    )
    gr.Markdown(LICENSE)

if __name__ == "__main__":
    demo.launch(server_port=8000, share=True)