metadata

language:
  - en
tags:
  - information retrieval
  - embedding model
  - visual information retrieval
metrics:
  - recall
pipeline_tag: feature-extraction

Memex: An OCR-free Visual-Based Document Embedding Model Based on MiniCPM-V-2.0

With MiniCPM-Visual-Embedding, it is possible to directly build knowledge base with raw PDF/Book/Document without any OCR technique nor OCR pipeline. The model only takes images as document-side inputs and produce vectors representing document pages. minicpm-visual-embedding-v0 is trained with over 30k paired query - visual document pages, including textual document, visual document, arxiv figures, industry documents, textbooks, ebooks, etc. The performance of minicpm-visual-embedding-v0 is on a par with a text embedding on text-oriented documents, and an advantages on visually-intensive documents.

Github Repo

News

2024-06-27: We released our first visual embedding model checkpoint minicpm-visual-embedding-v0 on huggingface.
2024-05-08: We committed our training code (full-parameter tuning with GradCache and DeepSpeed, supports large batch size across multiple GPUs with zero-stage1) and eval code.

Get started

Pip install all dependencies:

Pillow==10.1.0
timm==0.9.10
torch==2.1.2
torchvision==0.16.2
transformers==4.36.0
sentencepiece==0.1.99
numpy==1.26.0

First you are suggested to git clone this huggingface repo or download repo with huggingface_cli.

git lfs install
git clone https://huggingface.co/RhapsodyAI/minicpm-visual-embedding-v0

huggingface-cli download RhapsodyAI/minicpm-visual-embedding-v0

from transformers import AutoModel
from transformers import AutoTokenizer
from PIL import Image
import torch

device = 'cuda:0'

# This function is borrowed from https://huggingface.co/intfloat/e5-mistral-7b-instruct
def last_token_pool(last_hidden_states, attention_mask):
    left_padding = (attention_mask[:, -1].sum() == attention_mask.shape[0])
    if left_padding:
        return last_hidden_states[:, -1]
    else:
        sequence_lengths = attention_mask.sum(dim=1) - 1
        batch_size = last_hidden_states.shape[0]
        return last_hidden_states[torch.arange(batch_size, device=last_hidden_states.device), sequence_lengths]

# Load model, be sure to substitute `model_path` by your model path 
model_path = '/local/path/to/model'
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
model = AutoModel.from_pretrained(model_path, trust_remote_code=True)
model.to(device)

# Load image to PIL.Image object
image_1 = Image.open('/local/path/to/images/memex.png').convert('RGB')
image_2 = Image.open('/local/path/to/images/us2020.png').convert('RGB')
image_3 = Image.open('/local/path/to/images/hard_negative.png').convert('RGB')

# User query
query_instruction = 'Represent this query for retrieving relavant document: '
query = 'Who was elected as president of United States in 2020?'
query_full = query_instruction + query

# Embed image documents
with torch.no_grad():
    p_outputs = model(text=['', '', ''], image=[image_1, image_2, image_3], tokenizer=tokenizer)
    p_reps = last_token_pool(p_outputs.last_hidden_state, p_outputs.attention_mask)

# Embed text queries
with torch.no_grad():
    q_outputs = model(text=[query_full], image=[None], tokenizer=tokenizer) # [B, s, d]
    q_reps = last_token_pool(q_outputs.last_hidden_state, q_outputs.attention_mask) # [B, d]

# Calculate similarities
scores = torch.matmul(q_reps, p_reps.T)
print(scores)

# tensor([[0.6506, 4.9630, 3.8614]], device='cuda:0')

Limitations

Currently, please ensure that dpi of input images be a high value like 300 dpi, a lower dpi like 100 may cause the model performance degrade. We will augment data and fix this in our latest version.