How to Build an End-to-End OCR Pipeline with Baidu’s Unlimited-OCR for High-Resolution Images and Multi-Page PDF Parsing

Baidu's Unlimited-OCR, a 3B-parameter vision-language model, enables end-to-end OCR for high-resolution images and multi-page PDFs without separate layout analysis, according to a tutorial published on Hugging Face. The tutorial demonstrates a complete pipeline using the model's tiled Gundam and Base inference modes, supporting structured output for dense layouts, tables, and cross-page content.

In this tutorial, we build a complete workflow for running Baidu’s Unlimited-OCR https://huggingface.co/baidu/Unlimited-OCR model on document images and multi-page PDFs. We configure the GPU environment, install the required dependencies, load the 3B-parameter vision-language model with automatic selection of bfloat16 or float16, and generate structured sample documents for testing. We then evaluate both the tiled Gundam inference mode and the faster Base mode for single-page OCR before extending the pipeline to multi-page PDF parsing with PyMuPDF and infer multi . Throughout the workflow, we preserve long-context generation settings, repetition controls, and structured output handling to process dense layouts, tables, paragraphs, and cross-page content in a reproducible end-to-end pipeline. python import subprocess, sys def pip install pkgs : subprocess.check call sys.executable, "-m", "pip", "install", "-q", pkgs print " Installing dependencies 1-2 min ..." pip install "transformers==4.57.1", "Pillow", "matplotlib", "einops", "addict", "easydict", "pymupdf", "psutil", "accelerate", print " Done." import os import torch from transformers import AutoModel, AutoTokenizer assert torch.cuda.is available , "No GPU detected In Colab: Runtime - Change runtime type - GPU." gpu name = torch.cuda.get device name 0 print f" GPU: {gpu name}" use bf16 = torch.cuda.is bf16 supported DTYPE = torch.bfloat16 if use bf16 else torch.float16 print f" Using dtype: {DTYPE}" MODEL NAME = "baidu/Unlimited-OCR" print " Downloading model ~6 GB for 3B params in BF16 . First run takes a while..." tokenizer = AutoTokenizer.from pretrained MODEL NAME, trust remote code=True model = AutoModel.from pretrained MODEL NAME, trust remote code=True, use safetensors=True, torch dtype=DTYPE, model = model.eval .cuda print " Model loaded and moved to GPU." We install the required libraries and prepare the Google Colab environment for Unlimited-OCR inference. We verify that a CUDA-enabled GPU is available and automatically choose bfloat16 or float16 based on hardware support. We then load the tokenizer and the 3B-parameter model from Hugging Face, switch them to evaluation mode, and move them to the GPU. python from PIL import Image, ImageDraw, ImageFont import textwrap os.makedirs "inputs", exist ok=True os.makedirs "outputs/single gundam", exist ok=True os.makedirs "outputs/single base", exist ok=True os.makedirs "outputs/multi page", exist ok=True def load font size : for path in "/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf", "/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf", : if os.path.exists path : return ImageFont.truetype path, size return ImageFont.load default def make sample page path, page no : W, H = 1240, 1754 img = Image.new "RGB", W, H , "white" d = ImageDraw.Draw img title f, head f, body f = load font 48 , load font 34 , load font 26 d.text 80, 70 , f"Quarterly Operations Report — Page {page no}", fill="black", font=title f d.line 80, 145 , W - 80, 145 , fill="black", width=3 body = "This document demonstrates Unlimited-OCR's one-shot long-horizon " "parsing. The model reads an entire page — headings, paragraphs, " "and tables — and emits structured text in a single decoding pass. " "Unlike classic OCR pipelines, no separate layout-analysis stage " "is required." y = 190 for line in textwrap.wrap body, width=72 : d.text 80, y , line, fill="black", font=body f y += 40 y += 30 d.text 80, y , f"Table {page no}: Regional Revenue USD, millions ", fill="black", font=head f y += 60 rows = "Region", "Q1", "Q2", "Q3" , "North", "12.4", "13.1", "15.0" , "South", "9.8", "10.2", "11.7" , "East", "14.3", "13.9", "16.2" , "West", "11.1", "12.5", "12.9" , col w, row h, x0 = 260, 56, 80 for r, row in enumerate rows : for c, cell in enumerate row : x = x0 + c col w d.rectangle x, y, x + col w, y + row h , outline="black", width=2 d.text x + 14, y + 12 , cell, fill="black", font=body f y += row h y += 50 footer = f"Note {page no}: Figures are illustrative. Multi-page mode stitches " "context across pages, so cross-page references remain coherent." for line in textwrap.wrap footer, width=72 : d.text 80, y , line, fill="black", font=body f y += 40 img.save path return path IMAGE PATH = make sample page "inputs/sample page 1.png", 1 PAGE 2 = make sample page "inputs/sample page 2.png", 2 PAGE 3 = make sample page "inputs/sample page 3.png", 3 print f" Sample pages written: {IMAGE PATH}, {PAGE 2}, {PAGE 3}" import matplotlib.pyplot as plt plt.figure figsize= 6, 8 plt.imshow Image.open IMAGE PATH plt.axis "off" plt.title "Input document page 1 " plt.show We create the required input and output directories and generate three realistic sample document pages with PIL. We add headings, paragraphs, tables, and footnotes to test the model on structured, layout-rich content. We also preview the first generated page with Matplotlib before sending it to the OCR pipeline. print "\n" + "=" 76 print "STEP 4: Single image — GUNDAM mode tiled, high detail " print "=" 76 model.infer tokenizer, prompt="<image document parsing.", image file=IMAGE PATH, output path="outputs/single gundam", base size=1024, image size=640, crop mode=True, max length=32768, no repeat ngram size=35, ngram window=128, save results=True, We run single-image OCR using Gundam mode, which combines a global document view with tiled image crops. We enable crop mode and use a smaller tile size to preserve fine text and improve recognition on dense document layouts. We also configure long-output generation and repetition controls to ensure the model produces stable, structured results. print "\n" + "=" 76 print "STEP 5: Single image — BASE mode single view, faster " print "=" 76 model.infer tokenizer, prompt="<image document parsing.", image file=IMAGE PATH, output path="outputs/single base", base size=1024, image size=1024, crop mode=False, max length=32768, no repeat ngram size=35, ngram window=128, save results=True, We process the same document using Base mode with a single 1024-pixel image view. We turn off image cropping to reduce inference complexity and improve processing speed for clean, clearly printed pages. We retain the same output length and repetition-control settings to directly compare Base mode with Gundam mode. print "\n" + "=" 76 print "STEP 6: Multi-page / PDF parsing" print "=" 76 import tempfile import fitz def pdf to images pdf path, dpi=300 : """Rasterize every PDF page to a PNG; return the list of image paths.""" doc = fitz.open pdf path tmp dir = tempfile.mkdtemp prefix="pdf ocr " mat = fitz.Matrix dpi / 72, dpi / 72 paths = for i, page in enumerate doc : out = os.path.join tmp dir, f"page {i + 1:04d}.png" page.get pixmap matrix=mat .save out paths.append out doc.close return paths SAMPLE PDF = "inputs/sample doc.pdf" pdf = fitz.open for p in IMAGE PATH, PAGE 2, PAGE 3 : img doc = fitz.open p rect = img doc 0 .rect pdf bytes = img doc.convert to pdf img pdf = fitz.open "pdf", pdf bytes page = pdf.new page width=rect.width, height=rect.height page.show pdf page rect, img pdf, 0 pdf.save SAMPLE PDF pdf.close print f" Built sample PDF: {SAMPLE PDF}" page images = pdf to images SAMPLE PDF, dpi=300 print f" Rasterized {len page images } pages" model.infer multi tokenizer, prompt="<image Multi page parsing.", image files=page images, output path="outputs/multi page", image size=1024, max length=32768, no repeat ngram size=35, ngram window=1024, save results=True, We create a three-page PDF from the generated document images and rasterize each page of the PDF into a high-resolution PNG using PyMuPDF. We pass the resulting page-image sequence to infer multi so that the model can parse the complete document in a single long-horizon inference operation. We also widen the n-gram repetition window to maintain stable decoding across multiple pages. print "\n" + "=" 76 print "STEP 7: Saved outputs" print "=" 76 TEXT EXTS = {".txt", ".md", ".mmd", ".json"} def show outputs root : print f"\n--- {root} ---" if not os.path.isdir root : print " no output directory found " return for dirpath, , files in os.walk root : for fn in sorted files : fp = os.path.join dirpath, fn size = os.path.getsize fp print f" {fp} {size:,} bytes " if os.path.splitext fn 1 .lower in TEXT EXTS: with open fp, "r", encoding="utf-8", errors="replace" as f: content = f.read preview = content :1500 print " " + "-" 60 print "\n".join " | " + ln for ln in preview.splitlines if len content 1500: print f" | ... {len content - 1500:,} more chars " print " " + "-" 60 for out dir in "outputs/single gundam", "outputs/single base", "outputs/multi page" : show outputs out dir print """ ============================================================================ DONE — CHEAT SHEET ============================================================================ Single image, dense/small text .... infer , gundam 640 + crop mode=True Single image, clean print ......... infer , base 1024, crop mode=False Multi-page or PDF ................. infer multi , image size=1024, ngram window=1024 Long documents .................... keep max length=32768 and the no repeat ngram settings — they prevent degeneration on long outputs. Your own files .................... upload via Colab sidebar, point image file / pdf to images at them. ============================================================================ """ We inspect the output directories created by the single-page and multi-page inference runs. We list every generated file and display previews of supported text, Markdown, MMD, and JSON artifacts. We conclude the workflow with a concise reference summarizing the recommended inference modes for dense images, clean pages, and multi-page PDFs. In conclusion, we completed a practical OCR pipeline that handles both high-detail single-page documents and long multi-page PDFs within Google Colab. We compared Gundam and Base inference modes, rasterized PDFs into model-ready page images, ran long-horizon document parsing, and inspected the generated text, Markdown, and auxiliary artifacts directly from the output directories. We also configured the workflow to adapt to different GPU capabilities while retaining the generation parameters required for stable long-document decoding. It provides us with a reusable foundation for applying Unlimited-OCR to reports, scanned forms, technical documents, tables, and other layout-rich content without relying on a separate traditional OCR and layout-analysis stack. Check out the Full Code here. Also, feel free to follow us on and don’t forget to join our Twitter https://x.com/intent/follow?screen name=marktechpost and Subscribe to 150k+ML SubReddit https://www.reddit.com/r/machinelearningnews/ . Wait are you on telegram? our Newsletter https://www.aidevsignals.com/ now you can join us on telegram as well. https://t.me/machinelearningresearchnews Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us https://forms.gle/wbash1wF6efRj8G58 Sana Hassan, a consulting intern at Marktechpost and dual-degree student at IIT Madras, is passionate about applying technology and AI to address real-world challenges. With a keen interest in solving practical problems, he brings a fresh perspective to the intersection of AI and real-life solutions.