Skip to content
Other

PP-DocLayoutV3

document layout analysis module

2 min readupdated

This model analyzes document structure, even on complex or distorted surfaces like screen photos and crumpled paper. Unlike standard methods, it predicts multi-point bounding polygons instead of simple rectangles and determines the logical reading order in a single pass. Built as a core component for PaddleOCR-VL and GLM-OCR, it is aimed at developers and researchers working on high-precision text recognition and real-world document parsing.

Key features

  • Predicts multi-point bounding boxes for layout elements instead of standard two-point boxes.
  • Determines logical reading order for curved and skewed surfaces in a single pass.
  • Cuts down on cascading errors when processing complex, non-planar document images.
  • Integrates with modern vision-language systems like PaddleOCR-VL.

What you can do

  • Run the model via the transformers library to analyze the structure and layout of scanned documents.
  • Extract text in the proper logical reading order from photos of tilted pages or curved surfaces.
  • Plug it into document processing pipelines to accurately detect tables, headings, and paragraphs.

Example

import requests
from PIL import Image
from transformers import AutoImageProcessor, AutoModelForObjectDetection

model_path = "123explorer/pp-doclayoutv3"
model = AutoModelForObjectDetection.from_pretrained(model_path)
image_processor = AutoImageProcessor.from_pretrained(model_path)

image = Image.open(requests.get("[https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/layout_demo.jpg](https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/layout_demo.jpg)", stream=True).raw)
inputs = image_processor(images=image, return_tensors="pt")

outputs = model(**inputs)
results = image_processor.post_process_object_detection(outputs, target_sizes=[image.size[::-1]])
Read moreDetails
Library: transformers
Task: object-detection
Accepted at ECCV 2026
Supports multilingual documents
Examples
Discuss Telegram 0 28

Will it run?

Runs in F32

Fits without quantisation.

  • F32weights 0.1 GB · ready files0.6 GB
  • BF16weights 0.1 GB0.6 GB
  • FP8 / INT8recommendedweights 0.0 GB0.5 GB
  • 4-bitweights 0.0 GB0.5 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · F32 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # 123explorer/pp-doclayoutv3
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download 123explorer/pp-doclayoutv3 --local-dir ~/models/pp-doclayoutv3
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to pp-doclayoutv3
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of PP-DocLayoutV3 (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 4All on Hugging Face
Fine-tunes 3All on Hugging Face

About the model

123explorer/pp-doclayoutv3
Author
123explorer
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

280

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments