app.py

import time
from threading import Thread

import gradio as gr
import torch
from PIL import Image
from transformers import AutoProcessor, LlavaForConditionalGeneration, TextIteratorStreamer, TextStreamer

import spaces
import argparse

from llava_llama3.model.builder import load_pretrained_model
from llava_llama3.constants import IMAGE_TOKEN_INDEX, DEFAULT_IMAGE_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN
from llava_llama3.conversation import conv_templates, SeparatorStyle
from llava_llama3.utils import disable_torch_init
from llava_llama3.mm_utils import process_images, tokenizer_image_token, get_model_name_from_path
from llava_llama3.serve.cli import chat_llava

import requests
from io import BytesIO
import base64
import os
import glob
import pandas as pd
from tqdm import tqdm
import json

root_path = os.path.dirname(os.path.abspath(__file__))
print(root_path)
os.environ['GRADIO_TEMP_DIR'] = root_path

parser = argparse.ArgumentParser()
parser.add_argument("--model-path", type=str, default="TheFinAI/FinLLaVA")
parser.add_argument("--device", type=str, default="cuda")
parser.add_argument("--conv-mode", type=str, default="llama_3")
parser.add_argument("--temperature", type=float, default=0.01)
parser.add_argument("--max-new-tokens", type=int, default=512)
parser.add_argument("--load-8bit", action="store_true")
parser.add_argument("--load-4bit", action="store_true")
args = parser.parse_args()

args.load_8bit = True

# Load model
tokenizer, llava_model, image_processor, context_len = load_pretrained_model(
    args.model_path, 
    None, 
    'llava_llama3', 
    args.load_8bit, 
    args.load_4bit, 
    device=args.device)


def bot_streaming(message, history):
    print ("triggered")
    print(message)
    image_file = None
    if message["files"]:
        if type(message["files"][-1]) == dict:
            image_file = message["files"][-1]["path"]
        else:
            image_file = message["files"][-1]
    else:
        for hist in history:
            if type(hist[0]) == tuple:
                image_file = hist[0][0]
                
    if image_file is None:
        gr.Error("You need to upload an image for LLaVA to work.")
        return
    
    streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)

    @spaces.GPU
    def generate():
        print('Running chat')
        output = chat_llava(
                    args=args,
                    image_file=image_file,
                    text=message['text'],
                    tokenizer=tokenizer,
                    model=llava_model,
                    image_processor=image_processor,
                    context_len=context_len,
                    streamer=streamer)
        return output

    thread = Thread(target=generate)
    thread.start()
    # thread.join()

    buffer = ""
    # output = generate()
    for new_text in streamer:
        buffer += new_text
        generated_text_without_prompt = buffer
        time.sleep(0.06)
        print (generated_text_without_prompt)
        yield generated_text_without_prompt

chatbot = gr.Chatbot(scale=1)
chat_input = gr.MultimodalTextbox(interactive=True, file_types=["image"], placeholder="Enter message or upload file...", show_label=False)
with gr.Blocks(fill_height=True) as demo:
    gr.ChatInterface(
        fn=bot_streaming,
        title="FinLLaVA Demo",
        examples=[
            {"text": "What is in this picture?", "files": ["http://images.cocodataset.org/val2017/000000039769.jpg"]},
            {"text": "What is the spending on Healthcare in July? A. 450 B. 600 C. 520 D. 510", "files": ["image_107.png"]},
            {"text": "If 2012 net periodic opeb cost increased at the same pace as the pension cost, what would the estimated 2013 cost be in millions? A. 14.83333 B. 12.5 C. 15.5 D. 13.5", "files": ["image_659.png"]},
        ],
        description="",
        stop_btn="Stop Generation",
        multimodal=True,
        textbox=chat_input,
        chatbot=chatbot,
    )

demo.queue(api_open=False)
demo.launch(show_api=False, share=False)