layoutlmv3 docvqa t11c5000

Providerxhyi
Categorydocument-question-answering
LicenseApache-2.0
Downloads168
Stars0

Overview

LayoutLMv3 is a multimodal transformer designed for Document Visual Question Answering (DocVQA). Unlike traditional OCR-based pipelines that flatten text, this model treats text, layout coordinates, and image pixels as unified tokens. For developers, this means it can reason over the spatial relationship between elements—such as identifying a value based on its proximity to a specific label in a form or table. It is particularly effective for automating data extraction from invoices, receipts, and structured reports where visual context is critical for accuracy. The model integrates well into document processing pipelines and offers a significant performance boost over text-only models when handling complex layouts.

Highlights

  • Unified multimodal processing of text, layout, and images
  • High accuracy for spatial reasoning in structured documents
  • Ideal for automated invoice and form data extraction
  • Apache-2.0 license ensures flexible commercial integration

Usage

Install
# Install Hugging Face transformers
pip install transformers torch
SDK Usage
# Load model with transformers
from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained("xhyi/layoutlmv3_docvqa_t11c5000")
tokenizer = AutoTokenizer.from_pretrained("xhyi/layoutlmv3_docvqa_t11c5000")

Hugging Face Download

We recommend downloading the model via the Hugging Face CLI or Hub SDK.

Guidance：Before downloading, install huggingface_hub with:

Guidance 
pip install -U huggingface_hub

CLI Download

Download the full repository

Download the full repository 
huggingface-cli download xhyi/layoutlmv3_docvqa_t11c5000

Download a single file to a local folder (e.g. config.json into ./dir)

Download a single file to a local folder (e.g. config.json into ./dir) 
huggingface-cli download xhyi/layoutlmv3_docvqa_t11c5000 config.json --local-dir ./dir

See the official docs for more CLI options

SDK Download

SDK Download 
# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('xhyi/layoutlmv3_docvqa_t11c5000')

Git Download

Make sure git-lfs is installed first

Git Download 
git lfs install
git clone https://huggingface.co/xhyi/layoutlmv3_docvqa_t11c5000

To skip LFS large-file downloads, use:

Skip LFS 
GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/xhyi/layoutlmv3_docvqa_t11c5000

Model files are hosted on the Hugging Face Hub — download directly via HF CLI / SDK / Git, not through this site.

PyTorch / Transformers Usage

Install Transformers

Install Transformers 
pip install -U transformers torch

Load the model and run inference

Load the model and run inference 
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained('xhyi/layoutlmv3_docvqa_t11c5000')
tokenizer = AutoTokenizer.from_pretrained('xhyi/layoutlmv3_docvqa_t11c5000')

Full Documentation

来源: HuggingFace

LayoutLMv3: DocVQA Replication WIP

See experiments code: <https://github.com/redthing1/layoutlm_experiments>

View official source
Join our Telegram