RAG multimodale con Milvus

Se vuoi vedere il risultato finale di questo tutorial, puoi andare direttamente alla demo online.

Questo tutorial illustra il RAG multimodale basato su Milvus, il modello BGE visualizzato e GPT-4o. Con questo sistema, gli utenti possono caricare un'immagine e modificare le istruzioni testuali, che vengono elaborate dal modello di recupero composito di BGE per cercare immagini candidate. GPT-4o funge quindi da sistema di riclassificazione, selezionando l’immagine più adatta e fornendo la motivazione alla base della scelta. Questa potente combinazione consente un’esperienza di ricerca delle immagini fluida e intuitiva, sfruttando Milvus per un recupero efficiente, il modello BGE per un’elaborazione e un abbinamento precisi delle immagini e GPT-4o per una riclassificazione avanzata.

Preparazione

Installazione delle dipendenze

$ pip install --upgrade pymilvus openai datasets opencv-python timm einops ftfy peft tqdm

$ git clone https://github.com/FlagOpen/FlagEmbedding.git
$ pip install -e FlagEmbedding

Scaricare i dati

Il comando seguente scaricherà i dati di esempio ed li estrarrà in una cartella locale denominata “./images_folder”, che include:

• images: un sottoinsieme di Amazon Reviews 2023 contenente circa 900 immagini delle categorie “Appliance”, “Cell_Phones_and_Accessories” ed “Electronics”.
• leopard.jpg: un'immagine di esempio per la query.

$ wget https://github.com/milvus-io/bootcamp/releases/download/data/amazon_reviews_2023_subset.tar.gz
$ tar -xzf amazon_reviews_2023_subset.tar.gz

Caricare il modello di embedding

Utilizzeremo il modello BGE visualizzato “bge-visualized-base-en-v1.5” per generare gli embedding sia per le immagini che per il testo.

1. Scaricare i pesi

$ wget https://huggingface.co/BAAI/bge-visualized/resolve/main/Visualized_base_en_v1.5.pth

2. Costruire l’encoder

import torch
from visual_bge.modeling import Visualized_BGE


class Encoder:
    def __init__(self, model_name: str, model_path: str):
        self.model = Visualized_BGE(model_name_bge=model_name, model_weight=model_path)
        self.model.eval()

    def encode_query(self, image_path: str, text: str) -> list[float]:
        with torch.no_grad():
            query_emb = self.model.encode(image=image_path, text=text)
        return query_emb.tolist()[0]

    def encode_image(self, image_path: str) -> list[float]:
        with torch.no_grad():
            query_emb = self.model.encode(image=image_path)
        return query_emb.tolist()[0]


model_name = "BAAI/bge-base-en-v1.5"
model_path = "./Visualized_base_en_v1.5.pth"  # Change to your own value if using a different model path
encoder = Encoder(model_name, model_path)

Caricare i dati

In questa sezione verranno caricate nel database delle immagini di esempio con i relativi embedding.

Generare gli embedding

Carica tutte le immagini JPEG dalla directory dei dati e applica l'encoder per convertire le immagini in embedding.

import os
from tqdm import tqdm
from glob import glob


# Generate embeddings for the image dataset
data_dir = (
    "./images_folder"  # Change to your own value if using a different data directory
)
image_list = glob(
    os.path.join(data_dir, "images", "*.jpg")
)  # We will only use images ending with ".jpg"
image_dict = {}
for image_path in tqdm(image_list, desc="Generating image embeddings: "):
    try:
        image_dict[image_path] = encoder.encode_image(image_path)
    except Exception as e:
        print(f"Failed to generate embedding for {image_path}. Skipped.")
        continue
print("Number of encoded images:", len(image_dict))

Generating image embeddings: 100%|██████████| 900/900 [00:20<00:00, 44.08it/s]

Number of encoded images: 900

Inserimento in Milvus

Inserire le immagini con i percorsi e gli embedding corrispondenti nella collezione di Milvus.

from pymilvus import MilvusClient


dim = len(list(image_dict.values())[0])
collection_name = "multimodal_rag_demo"

# Connect to Milvus client given URI
milvus_client = MilvusClient(uri="./milvus_demo.db")

# Create Milvus Collection
# By default, vector field name is "vector"
milvus_client.create_collection(
    collection_name=collection_name,
    auto_id=True,
    dimension=dim,
    enable_dynamic_field=True,
)

# Insert data into collection
milvus_client.insert(
    collection_name=collection_name,
    data=[{"image_path": k, "vector": v} for k, v in image_dict.items()],
)

DEBUG:pymilvus.milvus_client.milvus_client:Created new connection using: 7f33daeed99a4d8e8a5e28d47673ecc8
DEBUG:pymilvus.milvus_client.milvus_client:Successfully created collection: multimodal_rag_demo
DEBUG:pymilvus.milvus_client.milvus_client:Successfully created an index on collection: multimodal_rag_demo





{'insert_count': 900,
 'ids': [451537887696781312, 451537887696781313, ..., 451537887696782211],
 'cost': 0}

Ricerca multimodale con Generative Reranker

In questa sezione, cercheremo innanzitutto le immagini pertinenti tramite una query multimodale, quindi utilizzeremo il servizio LLM per riorganizzare i risultati e individuare quello migliore con relativa spiegazione.

Eseguire la ricerca

Ora siamo pronti per eseguire la ricerca avanzata di immagini con dati di query composti sia da immagini che da istruzioni testuali.

query_image = os.path.join(
    data_dir, "leopard.jpg"
)  # Change to your own query image path
query_text = "phone case with this image theme"

# Generate query embedding given image and text instructions
query_vec = encoder.encode_query(image_path=query_image, text=query_text)

search_results = milvus_client.search(
    collection_name=collection_name,
    data=[query_vec],
    output_fields=["image_path"],
    limit=9,  # Max number of search results to return
    search_params={"metric_type": "COSINE", "params": {}},  # Search parameters
)[0]

retrieved_images = [hit.get("entity").get("image_path") for hit in search_results]
print(retrieved_images)

['./images_folder/images/518Gj1WQ-RL._AC_.jpg', './images_folder/images/41n00AOfWhL._AC_.jpg', './images_folder/images/51Wqge9HySL._AC_.jpg', './images_folder/images/51R2SZiywnL._AC_.jpg', './images_folder/images/516PebbMAcL._AC_.jpg', './images_folder/images/51RrgfYKUfL._AC_.jpg', './images_folder/images/515DzQVKKwL._AC_.jpg', './images_folder/images/51BsgVw6RhL._AC_.jpg', './images_folder/images/51INtcXu9FL._AC_.jpg']

Riorganizzazione della classifica con GPT-4o

Utilizzeremo un LLM per classificare le immagini e generare una spiegazione per il risultato migliore in base alla query dell’utente e ai risultati recuperati.

1. Creare una vista panoramica

import numpy as np
import cv2

img_height = 300
img_width = 300
row_count = 3


def create_panoramic_view(query_image_path: str, retrieved_images: list) -> np.ndarray:
    """
    creates a 5x5 panoramic view image from a list of images

    args:
        images: list of images to be combined

    returns:
        np.ndarray: the panoramic view image
    """
    panoramic_width = img_width * row_count
    panoramic_height = img_height * row_count
    panoramic_image = np.full(
        (panoramic_height, panoramic_width, 3), 255, dtype=np.uint8
    )

    # create and resize the query image with a blue border
    query_image_null = np.full((panoramic_height, img_width, 3), 255, dtype=np.uint8)
    query_image = Image.open(query_image_path).convert("RGB")
    query_array = np.array(query_image)[:, :, ::-1]
    resized_image = cv2.resize(query_array, (img_width, img_height))

    border_size = 10
    blue = (255, 0, 0)  # blue color in BGR
    bordered_query_image = cv2.copyMakeBorder(
        resized_image,
        border_size,
        border_size,
        border_size,
        border_size,
        cv2.BORDER_CONSTANT,
        value=blue,
    )

    query_image_null[img_height * 2 : img_height * 3, 0:img_width] = cv2.resize(
        bordered_query_image, (img_width, img_height)
    )

    # add text "query" below the query image
    text = "query"
    font_scale = 1
    font_thickness = 2
    text_org = (10, img_height * 3 + 30)
    cv2.putText(
        query_image_null,
        text,
        text_org,
        cv2.FONT_HERSHEY_SIMPLEX,
        font_scale,
        blue,
        font_thickness,
        cv2.LINE_AA,
    )

    # combine the rest of the images into the panoramic view
    retrieved_imgs = [
        np.array(Image.open(img).convert("RGB"))[:, :, ::-1] for img in retrieved_images
    ]
    for i, image in enumerate(retrieved_imgs):
        image = cv2.resize(image, (img_width - 4, img_height - 4))
        row = i // row_count
        col = i % row_count
        start_row = row * img_height
        start_col = col * img_width

        border_size = 2
        bordered_image = cv2.copyMakeBorder(
            image,
            border_size,
            border_size,
            border_size,
            border_size,
            cv2.BORDER_CONSTANT,
            value=(0, 0, 0),
        )
        panoramic_image[
            start_row : start_row + img_height, start_col : start_col + img_width
        ] = bordered_image

        # add red index numbers to each image
        text = str(i)
        org = (start_col + 50, start_row + 30)
        (font_width, font_height), baseline = cv2.getTextSize(
            text, cv2.FONT_HERSHEY_SIMPLEX, 1, 2
        )

        top_left = (org[0] - 48, start_row + 2)
        bottom_right = (org[0] - 48 + font_width + 5, org[1] + baseline + 5)

        cv2.rectangle(
            panoramic_image, top_left, bottom_right, (255, 255, 255), cv2.FILLED
        )
        cv2.putText(
            panoramic_image,
            text,
            (start_col + 10, start_row + 30),
            cv2.FONT_HERSHEY_SIMPLEX,
            1,
            (0, 0, 255),
            2,
            cv2.LINE_AA,
        )

    # combine the query image with the panoramic view
    panoramic_image = np.hstack([query_image_null, panoramic_image])
    return panoramic_image

Combina l’immagine della query e le immagini recuperate con gli indici in una vista panoramica.

from PIL import Image

combined_image_path = os.path.join(data_dir, "combined_image.jpg")
panoramic_image = create_panoramic_view(query_image, retrieved_images)
cv2.imwrite(combined_image_path, panoramic_image)

combined_image = Image.open(combined_image_path)
show_combined_image = combined_image.resize((300, 300))
show_combined_image.show()

Creazione di una vista panoramica

2. Riclassificare e fornire una spiegazione

Invieremo l'immagine combinata al servizio LLM multimodale insieme a prompt appropriati per classificare i risultati recuperati con una spiegazione. Per abilitare GPT-4o come LLM, è necessario preparare la propria chiave API OpenAI.

import requests
import base64

openai_api_key = "sk-***"  # Change to your OpenAI API Key


def generate_ranking_explanation(
    combined_image_path: str, caption: str, infos: dict = None
) -> tuple[list[int], str]:
    with open(combined_image_path, "rb") as image_file:
        base64_image = base64.b64encode(image_file.read()).decode("utf-8")

    information = (
        "You are responsible for ranking results for a Composed Image Retrieval. "
        "The user retrieves an image with an 'instruction' indicating their retrieval intent. "
        "For example, if the user queries a red car with the instruction 'change this car to blue,' a similar type of car in blue would be ranked higher in the results. "
        "Now you would receive instruction and query image with blue border. Every item has its red index number in its top left. Do not misunderstand it. "
        f"User instruction: {caption} \n\n"
    )

    # add additional information for each image
    if infos:
        for i, info in enumerate(infos["product"]):
            information += f"{i}. {info}\n"

    information += (
        "Provide a new ranked list of indices from most suitable to least suitable, followed by an explanation for the top 1 most suitable item only. "
        "The format of the response has to be 'Ranked list: []' with the indices in brackets as integers, followed by 'Reasons:' plus the explanation why this most fit user's query intent."
    )

    headers = {
        "Content-Type": "application/json",
        "Authorization": f"Bearer {openai_api_key}",
    }

    payload = {
        "model": "gpt-4o",
        "messages": [
            {
                "role": "user",
                "content": [
                    {"type": "text", "text": information},
                    {
                        "type": "image_url",
                        "image_url": {"url": f"data:image/jpeg;base64,{base64_image}"},
                    },
                ],
            }
        ],
        "max_tokens": 300,
    }

    response = requests.post(
        "https://api.openai.com/v1/chat/completions", headers=headers, json=payload
    )
    result = response.json()["choices"][0]["message"]["content"]

    # parse the ranked indices from the response
    start_idx = result.find("[")
    end_idx = result.find("]")
    ranked_indices_str = result[start_idx + 1 : end_idx].split(",")
    ranked_indices = [int(index.strip()) for index in ranked_indices_str]

    # extract explanation
    explanation = result[end_idx + 1 :].strip()

    return ranked_indices, explanation

Ottenere gli indici delle immagini dopo la classificazione e la motivazione del risultato migliore:

ranked_indices, explanation = generate_ranking_explanation(
    combined_image_path, query_text
)

3. Visualizza il risultato migliore con la spiegazione

print(explanation)

best_index = ranked_indices[0]
best_img = Image.open(retrieved_images[best_index])
best_img = best_img.resize((150, 150))
best_img.show()

Reasons: The most suitable item for the user's query intent is index 6 because the instruction specifies a phone case with the theme of the image, which is a leopard. The phone case with index 6 has a thematic design resembling the leopard pattern, making it the closest match to the user's request for a phone case with the image theme.

Il risultato migliore

Implementazione rapida

Per scoprire come avviare una demo online con questo tutorial, consulta l’applicazione di esempio.