In this tutorial, we build a complete workflow for running Baidu’s 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.
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.
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 Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.
Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us
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.
