app.py

from transformers import MllamaForConditionalGeneration, AutoProcessor, TextIteratorStreamer
from PIL import Image
import requests
import torch
from threading import Thread
import gradio as gr
from gradio import FileData
import time
import spaces
import os

# Load Hugging Face token from environment variables
hf_token = os.getenv("HF_AUTH_TOKEN")
if not hf_token:
    raise ValueError("Hugging Face token not found. Set HF_AUTH_TOKEN in your Space settings.")

# Model checkpoint
ckpt = "meta-llama/Llama-3.2-11B-Vision-Instruct"

# Load model and processor with authentication
model = MllamaForConditionalGeneration.from_pretrained(
    ckpt,
    torch_dtype=torch.bfloat16,
    token=hf_token
).to("cuda")

processor = AutoProcessor.from_pretrained(ckpt, token=hf_token)


@spaces.GPU
def bot_streaming(message, history, max_new_tokens=4500):
    txt = message["text"]
    ext_buffer = f"{txt}"
    
    messages = [] 
    images = []
    
    # Process conversation history
    for i, msg in enumerate(history): 
        if isinstance(msg[0], tuple):
            messages.append({"role": "user", "content": [{"type": "text", "text": history[i+1][0]}, {"type": "image"}]})
            messages.append({"role": "assistant", "content": [{"type": "text", "text": history[i+1][1]}]})
            images.append(Image.open(msg[0][0]).convert("RGB"))
        elif isinstance(history[i-1], tuple) and isinstance(msg[0], str):
            # Messages are already handled
            pass
        elif isinstance(history[i-1][0], str) and isinstance(msg[0], str):  # Text-only turn
            messages.append({"role": "user", "content": [{"type": "text", "text": msg[0]}]})
            messages.append({"role": "assistant", "content": [{"type": "text", "text": msg[1]}]})

    # Add current message
    if len(message["files"]) == 1:
        if isinstance(message["files"][0], str):  # Examples
            image = Image.open(message["files"][0]).convert("RGB")
        else:  # Regular input
            image = Image.open(message["files"][0]["path"]).convert("RGB")
        images.append(image)
        messages.append({"role": "user", "content": [{"type": "text", "text": txt}, {"type": "image"}]})
    else:
        messages.append({"role": "user", "content": [{"type": "text", "text": txt}]})

    # Prepare inputs
    texts = processor.apply_chat_template(messages, add_generation_prompt=True)
    if images == []:
        inputs = processor(text=texts, return_tensors="pt").to("cuda")
    else:
        inputs = processor(text=texts, images=images, return_tensors="pt").to("cuda")

    streamer = TextIteratorStreamer(processor, skip_special_tokens=True, skip_prompt=True)
    generation_kwargs = dict(inputs, streamer=streamer, max_new_tokens=max_new_tokens)
    generated_text = ""
    
    # Stream generation in a separate thread
    thread = Thread(target=model.generate, kwargs=generation_kwargs)
    thread.start()
    buffer = ""
    
    for new_text in streamer:
        buffer += new_text
        time.sleep(0.01)
        yield buffer


# Gradio Interface
demo = gr.ChatInterface(
    fn=bot_streaming,
    title="Multimodal Llama",
    examples=[],
    textbox=gr.MultimodalTextbox(),
    additional_inputs=[
        gr.Slider(
            minimum=10,
            maximum=5000,
            value=250,
            step=10,
            label="Maximum number of new tokens to generate",
        )
    ],
    cache_examples=False,
    description=(
        "Try Multimodal Llama by Meta with transformers in this demo. "
        "Upload an image, and start chatting about it, or simply try one of the examples below. "
        "To learn more about Llama Vision, visit [our blog post](https://huggingface.co./blog/llama32)."
    ),
    stop_btn="Stop Generation",
    fill_height=True,
    multimodal=True
)

demo.launch(debug=True)