import streamlit as st from text2image import get_model, get_tokenizer, get_image_transform from utils import text_encoder from torchvision import transforms from PIL import Image from jax import numpy as jnp import pandas as pd import numpy as np import requests import psutil import time import jax import gc preprocess = transforms.Compose( [ transforms.ToTensor(), transforms.Normalize( (0.48145466, 0.4578275, 0.40821073), (0.26862954, 0.26130258, 0.27577711) ), ] ) def pad_to_square(image, size=224): ratio = float(size) / max(image.size) new_size = tuple([int(x * ratio) for x in image.size]) image = image.resize(new_size, Image.ANTIALIAS) new_image = Image.new("RGB", size=(size, size), color=(128, 128, 128)) new_image.paste(image, ((size - new_size[0]) // 2, (size - new_size[1]) // 2)) return new_image def image_encoder(image, model): image = np.transpose(image, (0, 2, 3, 1)) features = model.get_image_features(image) features /= jnp.linalg.norm(features, keepdims=True) return features def gen_image_batch(image_url, image_size=224, pixel_size=10): n_pixels = image_size // pixel_size + 1 image_batch = [] masks = [] image_raw = requests.get(image_url, stream=True).raw image = Image.open(image_raw).convert("RGB") image = pad_to_square(image, size=image_size) gray = np.ones_like(image) * 128 mask = np.ones_like(image) image_batch.append(image) masks.append(mask) for i in range(0, n_pixels): for j in range(i + 1, n_pixels): m = mask.copy() m[: min(i * pixel_size, image_size) + 1, :] = 0 m[min(j * pixel_size, image_size) + 1 :, :] = 0 neg_m = 1 - m image_batch.append(image * m + gray * neg_m) masks.append(m) for i in range(0, n_pixels + 1): for j in range(i + 1, n_pixels + 1): m = mask.copy() m[:, : min(i * pixel_size + 1, image_size)] = 0 m[:, min(j * pixel_size + 1, image_size) :] = 0 neg_m = 1 - m image_batch.append(image * m + gray * neg_m) masks.append(m) return image_batch, masks def get_heatmap(image_url, text, pixel_size=10, iterations=3): tokenizer = get_tokenizer() model = get_model() image_size = model.config.vision_config.image_size text_embedding = text_encoder(text, model, tokenizer) images, masks = gen_image_batch( image_url, image_size=image_size, pixel_size=pixel_size ) input_image = images[0].copy() images = np.stack([preprocess(image) for image in images], axis=0) image_embeddings = jnp.asarray(image_encoder(images, model)) sims = [] scores = [] mask_val = jnp.zeros_like(masks[0]) for e, m in zip(image_embeddings, masks): sim = jnp.matmul(e, text_embedding.T) sims.append(sim) if len(sims) > 1: scores.append(sim * m) mask_val += 1 - m score = jnp.mean(jnp.clip(jnp.array(scores) - sims[0], 0, jnp.inf), axis=0) for i in range(iterations): score = jnp.clip(score - jnp.mean(score), 0, jnp.inf) score = (score - jnp.min(score)) / (jnp.max(score) - jnp.min(score)) return np.asarray(score), input_image def app(): st.title("Zero-Shot Localization") st.markdown( """ ### 👋 Ciao! Here you can find an example for zero-shot localization that will show you where in an image the model sees an object. The object location is computed by masking different areas of the image and looking at how the similarity to the image description changes. If you want to have a look at the implementation in detail, you can find it in [this Colab](https://colab.research.google.com/drive/10neENr1DEAFq_GzsLqBDo0gZ50hOhkOr?usp=sharing). On the two parameters: + the *pixel size* defines the resolution of the localization map. A pixel size of 15 means that 15 pixels in the original image will form 1 pixel in the heatmap. + The *refinement iterations* are just a cheap operation to reduce background noise. Too few iterations will leave a lot of noise. Too many will shrink the heatmap too much. 🤌 Italian mode on! 🤌 For example, try typing "gatto" (cat) or "cane" (dog) in the space for label and click "locate"! """ ) image_url = st.text_input( "You can input the URL of an image here...", value="https://www.tuttosuigatti.it/files/styles/full_width/public/images/featured/205/cani-e-gatti.jpg?itok=WAAiTGS6", ) MAX_ITER = 1 col1, col2 = st.columns([0.75, 0.25]) with col2: pixel_size = st.selectbox("Pixel Size", options=range(10, 26, 5), index=2) iterations = st.selectbox("Refinement Steps", options=range(3, 30, 3), index=0) compute = st.button("LOCATE") with col1: caption = st.text_input(f"Insert label...") if compute: with st.spinner("Waiting for resources..."): sleep_time = 5 while psutil.cpu_percent() > 50: time.sleep(sleep_time) if not caption or not image_url: st.error("Please choose one image and at least one label") else: with st.spinner( "Computing... This might take up to a few minutes depending on the current load 😕 \n" "Otherwise, you can use this [Colab notebook](https://colab.research.google.com/drive/10neENr1DEAFq_GzsLqBDo0gZ50hOhkOr?usp=sharing)" ): heatmap, image = get_heatmap(image_url, caption, pixel_size, iterations) with col1: st.image(image, use_column_width=True) st.image(heatmap, use_column_width=True) st.image(np.asarray(image) / 255.0 * heatmap, use_column_width=True) gc.collect() elif image_url: image = requests.get( image_url, stream=True, ).raw image = Image.open(image).convert("RGB") with col1: st.image(image)