app.py

import numpy as np
import torch
import torchvision.transforms as T
from PIL import Image
from torchvision.transforms.functional import InterpolationMode
from transformers import AutoModel, AutoTokenizer
import matplotlib.pyplot as plt
import random
import streamlit as st
import requests
from io import BytesIO


IMAGENET_MEAN = (0.485, 0.456, 0.406)
IMAGENET_STD = (0.229, 0.224, 0.225)

def build_transform(input_size):
    MEAN, STD = IMAGENET_MEAN, IMAGENET_STD
    transform = T.Compose([
        T.Lambda(lambda img: img.convert('RGB') if img.mode != 'RGB' else img),
        T.Resize((input_size, input_size), interpolation=InterpolationMode.BICUBIC),
        T.ToTensor(),
        T.Normalize(mean=MEAN, std=STD)
    ])
    return transform

def find_closest_aspect_ratio(aspect_ratio, target_ratios, width, height, image_size):
    best_ratio_diff = float('inf')
    best_ratio = (1, 1)
    area = width * height
    for ratio in target_ratios:
        target_aspect_ratio = ratio[0] / ratio[1]
        ratio_diff = abs(aspect_ratio - target_aspect_ratio)
        if ratio_diff < best_ratio_diff:
            best_ratio_diff = ratio_diff
            best_ratio = ratio
        elif ratio_diff == best_ratio_diff:
            if area > 0.5 * image_size * image_size * ratio[0] * ratio[1]:
                best_ratio = ratio
    return best_ratio

def dynamic_preprocess(image, min_num=1, max_num=12, image_size=448, use_thumbnail=False):
    orig_width, orig_height = image.size
    aspect_ratio = orig_width / orig_height

    # calculate the existing image aspect ratio
    target_ratios = set(
        (i, j) for n in range(min_num, max_num + 1) for i in range(1, n + 1) for j in range(1, n + 1) if
        i * j <= max_num and i * j >= min_num)
    target_ratios = sorted(target_ratios, key=lambda x: x[0] * x[1])

    # find the closest aspect ratio to the target
    target_aspect_ratio = find_closest_aspect_ratio(
        aspect_ratio, target_ratios, orig_width, orig_height, image_size)

    # calculate the target width and height
    target_width = image_size * target_aspect_ratio[0]
    target_height = image_size * target_aspect_ratio[1]
    blocks = target_aspect_ratio[0] * target_aspect_ratio[1]

    # resize the image
    resized_img = image.resize((target_width, target_height))
    processed_images = []
    for i in range(blocks):
        box = (
            (i % (target_width // image_size)) * image_size,
            (i // (target_width // image_size)) * image_size,
            ((i % (target_width // image_size)) + 1) * image_size,
            ((i // (target_width // image_size)) + 1) * image_size
        )
        # split the image
        split_img = resized_img.crop(box)
        processed_images.append(split_img)
    assert len(processed_images) == blocks
    if use_thumbnail and len(processed_images) != 1:
        thumbnail_img = image.resize((image_size, image_size))
        processed_images.append(thumbnail_img)
    return processed_images

def load_image(image_file, input_size=448, max_num=12):
    image = Image.open(image_file).convert('RGB')
    
    transform = build_transform(input_size=input_size)
    images = dynamic_preprocess(image, image_size=input_size, use_thumbnail=True, max_num=max_num)
    pixel_values = [transform(image) for image in images]
    pixel_values = torch.stack(pixel_values)
    return pixel_values

def prediction(model, image_file, question):
    question = f"<image>\n{question}"
    # set the max number of tiles in `max_num`
    pixel_values = load_image(image_file, max_num=12).to(torch.bfloat16).cuda()
    generation_config = dict(max_new_tokens=1024, do_sample=False)

    response = model.chat(tokenizer, pixel_values, question, generation_config)

    return response

# If you want to load a model using multiple GPUs, please refer to the `Multiple GPUs` section.
path = 'Ramji/slake_vqa_internvl_demo'
intern_model = AutoModel.from_pretrained(
    path,
    torch_dtype=torch.bfloat16,
    low_cpu_mem_usage=True,
    use_flash_attn=False,
    trust_remote_code=True).eval().cuda()
tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True, use_fast=False)

# Title of the Streamlit app
st.title("Image VQA")

# Step 1: Upload an image
st.header("Upload an Image")
uploaded_image = st.file_uploader("Choose an image...", type=["jpg", "png", "jpeg"])

# Step 2: Input a question
st.header("Ask a Question")
question = st.text_input("Type your question here:")

# Step 3: Handle the uploaded image by saving it and reading its path
if uploaded_image is not None:
    # Save the uploaded image to a file
    image_path = os.path.join("uploaded_images", uploaded_image.name)
    
    # Make sure the directory exists
    os.makedirs("uploaded_images", exist_ok=True)
    
    # Write the image to a file
    with open(image_path, "wb") as f:
        f.write(uploaded_image.getbuffer())
    
    # Read the image from the saved file path
    image = Image.open(image_path)
    
    # Display the uploaded image
    st.image(image, caption="Uploaded Image", use_column_width=True)
    
    st.write(f"Image saved at: {image_path}")

# Step 4: Display the typed question
if question:
    st.write(f"Your question: **{question}**")

# Optional: Process the image and question for a VLM (like CLIP or BLIP)
if uploaded_image and question:
    st.write("Processing the image and question...")

    output = prediction(intern_model, image_file, question)

    st.write("Model output: This is where the answer will appear.")