|
|
<!DOCTYPE html> |
|
|
|
|
|
<html lang="en" xmlns="http://www.w3.org/1999/html"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8"> |
|
|
|
|
|
<meta name="viewport" content="width=device-width, initial-scale=1"> |
|
|
<meta name="description" content="DoGaussian: Distributed-Oriented Gaussian Splatting for Large-Scale 3D Reconstruction Via Gaussian Consensus"> |
|
|
|
|
|
|
|
|
<meta name="twitter:card" content="summary_large_image"> |
|
|
<meta name="twitter:title" content="DoGaussian: Distributed-Oriented Gaussian Splatting for Large-Scale 3D Reconstruction Via Gaussian Consensus"> |
|
|
<meta name="twitter:description" content="NUS' DoGaussian is a distributed method that accelerates 3DGS training with multiple GPU nodes."> |
|
|
<meta name="twitter:image" content="https://aibluefisher.github.io/dogaussian/assets/social_card.png"> |
|
|
|
|
|
<title>DoGaussian</title> |
|
|
|
|
|
<link href="assets/bootstrap.min.css" rel="stylesheet"> |
|
|
<link rel="stylesheet" href="assets/all.min.css"> |
|
|
<link href="assets/style.css" rel="stylesheet"> |
|
|
|
|
|
<link rel="stylesheet" href="assets/slider.css"> |
|
|
|
|
|
<link rel="icon" href="assets/imgs/dog_icon.png"> |
|
|
<style type="text/css" abt="234"></style><script> |
|
|
var _countAA = 0 |
|
|
function doBBBd(){} |
|
|
doBBBd() |
|
|
document.addEventListener('keyup', function(){_countAA-=10;doBBBd()}, false) |
|
|
document.addEventListener('click', function(){_countAA-=10;doBBBd()}, false) |
|
|
|
|
|
</script></head> |
|
|
<body data-new-gr-c-s-check-loaded="14.1173.0" data-gr-ext-installed=""> |
|
|
|
|
|
<div class="container"> |
|
|
<div class="row text-center"> |
|
|
<h1 class="mt-3"><img src="assets/imgs/dog_icon.png" style="height: 50px; width: auto">DoGaussian: Distributed-Oriented Gaussian Splatting </h1> |
|
|
<h3 class="mb-4">for Large-Scale 3D Reconstruction Via Gaussian Consensus</h3> |
|
|
</div> |
|
|
<div class="row text-center"> |
|
|
<div class="col-sm-0 col-md-2"></div> |
|
|
<div class="col-sm-12 col-md-8"> |
|
|
<h6> |
|
|
<a href="https://aibluefisher.github.io/" target="_blank"><nobr>Yu Chen</nobr></a> |
|
|
<a href="https://www.comp.nus.edu.sg/~leegh/" target="_blank"><nobr>Gim Hee Lee</nobr></a> |
|
|
</h6> |
|
|
National University of Singapore |
|
|
</div> |
|
|
<div class="col-sm-0 col-md-2"></div> |
|
|
<div class="col-md-12 mt-3"> |
|
|
<a href="https://arxiv.org/abs/2405.13943" class="btn btn-secondary btn-sm" target="_blank"><i class="fa-solid fa-file-pdf"></i> arXiv</a> |
|
|
<a href="index.html" target="_blank" class="btn btn-secondary btn-sm"><i class="fa-brands fa-youtube"></i> Video</a> |
|
|
<a href="https://github.com/aibluefisher/DoGaussian" target="_blank" class="btn btn-secondary btn-sm"><i class="fa-brands fa-github"></i> Code</a> |
|
|
<a href="https://aibluefisher.github.io/dogaussian/#citation" class="btn btn-secondary btn-sm"><i class="fa-solid fa-file"></i> BibTeX</a> |
|
|
</div> |
|
|
</div> |
|
|
<div class="row"> |
|
|
<div class="col-sm-0 col-md-2"></div> |
|
|
<div class="col-sm-12 col-md-8"> |
|
|
|
|
|
<video class="w-100 mt-4" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/DoGaussian_demo_final.mp4" type="video/mp4"> |
|
|
</video> |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<h3 class="mt-3">Abstract</h3> |
|
|
<hr> |
|
|
<p> |
|
|
The recent advances in 3D Gaussian Splatting (3DGS) show promising results on the novel view |
|
|
synthesis (NVS) task. With its superior rendering performance and high-fidelity rendering quality, |
|
|
3DGS is excelling at its previous NeRF counterparts. The most recent 3DGS method focuses either on |
|
|
improving the instability of rendering efficiency or reducing the model size. On the other hand, |
|
|
the training efficiency of 3DGS on large-scale scenes has not gained much attention. In this work, |
|
|
we propose <b>DoGaussian</b>, a method that trains 3DGS distributedly. Our method first decomposes a scene into |
|
|
K blocks and then introduces the Alternating Direction Method of Multipliers (ADMM) into the training |
|
|
procedure of 3DGS. During training, our DoGaussian maintains one global 3DGS model on the master node and |
|
|
K local 3DGS models on the slave nodes. The K local 3DGS models are dropped after training and we only |
|
|
query the global 3DGS model during inference. The training time is reduced by scene decomposition, |
|
|
and the training convergence and stability are guaranteed through the consensus on the shared 3D Gaussians. |
|
|
Our method <b>accelerates the training of 3DGS by 6+ times</b> when evaluated on large-scale scenes while |
|
|
concurrently achieving state-of-the-art rendering quality. |
|
|
</p> |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<h3 class="mt-5">Watch <b>DoGaussian</b> reconstructed large-scale scenes</h3> |
|
|
<hr class="hr"> |
|
|
<p> |
|
|
We visualize the reconstruction results of <b>DoGaussian</b> on the Mill-19 and UrbanScene3D dataset. |
|
|
The results are recorded on <u>a browser of a MacBook (m1 chip with 8GB memory)</u> with freely moving |
|
|
camera trajectories that are highly different from the training views. |
|
|
<b>Use the controls to switch between scenes.</b> |
|
|
</p> |
|
|
|
|
|
<div id="DoGaussianResults" class="carousel slide" data-bs-ride="false"> |
|
|
<div class="carousel-indicators"> |
|
|
<button type="button" data-bs-target="#DoGaussianResults" data-bs-slide-to="0" class="active" aria-current="true" aria-label="Slide 1"></button> |
|
|
<button type="button" data-bs-target="#DoGaussianResults" data-bs-slide-to="1" aria-label="Slide 2"></button> |
|
|
<button type="button" data-bs-target="#DoGaussianResults" data-bs-slide-to="2" aria-label="Slide 3"></button> |
|
|
<button type="button" data-bs-target="#DoGaussianResults" data-bs-slide-to="3" aria-label="Slide 4"></button> |
|
|
<button type="button" data-bs-target="#DoGaussianResults" data-bs-slide-to="4" aria-label="Slide 5"></button> |
|
|
</div> |
|
|
<div class="carousel-inner"> |
|
|
<div class="carousel-item active"> |
|
|
<video class="w-100" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/building_dogaussian.mp4" type="video/mp4"> |
|
|
</video> |
|
|
<div class="carousel-caption"> |
|
|
<h5>Mill19 - Building</h5> |
|
|
</div> |
|
|
</div> |
|
|
<div class="carousel-item"> |
|
|
<video class="w-100" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/rubble_dogaussian.mp4" type="video/mp4"> |
|
|
</video> |
|
|
<div class="carousel-caption"> |
|
|
<h5>Mill-19 - Rubble</h5> |
|
|
</div> |
|
|
</div> |
|
|
<div class="carousel-item"> |
|
|
<video class="w-100" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/campus_dogaussian.mp4" type="video/mp4"> |
|
|
</video> |
|
|
<div class="carousel-caption"> |
|
|
<h5>UrbanScene3D - Campus</h5> |
|
|
</div> |
|
|
</div> |
|
|
<div class="carousel-item"> |
|
|
<video class="w-100" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/sciart_dogaussian.mp4" type="video/mp4"> |
|
|
</video> |
|
|
<div class="carousel-caption"> |
|
|
<h5>UrbanScene3D - Sci-Art</h5> |
|
|
</div> |
|
|
</div> |
|
|
<div class="carousel-item"> |
|
|
<video class="w-100" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/residence_dogaussian.mp4" type="video/mp4"> |
|
|
</video> |
|
|
<div class="carousel-caption"> |
|
|
<h5>UrbanScene3D - Residence</h5> |
|
|
</div> |
|
|
</div> |
|
|
</div> |
|
|
<button class="carousel-control-prev" type="button" data-bs-target="#DoGaussianResults" data-bs-slide="prev"> |
|
|
<span class="carousel-control-prev-icon" aria-hidden="true"></span> |
|
|
<span class="visually-hidden">Previous</span> |
|
|
</button> |
|
|
<button class="carousel-control-next" type="button" data-bs-target="#DoGaussianResults" data-bs-slide="next"> |
|
|
<span class="carousel-control-next-icon" aria-hidden="true"></span> |
|
|
<span class="visually-hidden">Next</span> |
|
|
</button> |
|
|
</div> |
|
|
|
|
|
<h3 class="mt-5">Watch <b>DoGaussian</b> reconstructed 3D Gaussian Primitives</h3> |
|
|
<hr class="hr"> |
|
|
<p> |
|
|
We visualize the reconstructed 3D Gaussian primitives of <b>DoGaussian</b> on the Mill-19 and UrbanScene3D dataset. |
|
|
The results are recorded on <u>a browser of a MacBook (m1 chip with 8GB memory)</u> with freely moving |
|
|
camera trajectories that are highly different from the training views. |
|
|
<b>Use the controls to switch between scenes.</b> |
|
|
</p> |
|
|
|
|
|
<div id="DoGaussianPclResults" class="carousel slide" data-bs-ride="false"> |
|
|
<div class="carousel-indicators"> |
|
|
<button type="button" data-bs-target="#DoGaussianPclResults" data-bs-slide-to="0" class="active" aria-current="true" aria-label="Slide 1"></button> |
|
|
<button type="button" data-bs-target="#DoGaussianPclResults" data-bs-slide-to="1" aria-label="Slide 2"></button> |
|
|
<button type="button" data-bs-target="#DoGaussianPclResults" data-bs-slide-to="2" aria-label="Slide 3"></button> |
|
|
<button type="button" data-bs-target="#DoGaussianPclResults" data-bs-slide-to="3" aria-label="Slide 4"></button> |
|
|
<button type="button" data-bs-target="#DoGaussianPclResults" data-bs-slide-to="4" aria-label="Slide 5"></button> |
|
|
</div> |
|
|
<div class="carousel-inner"> |
|
|
<div class="carousel-item active"> |
|
|
<video class="w-100" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/building_dogaussian_pcl.mp4" type="video/mp4"> |
|
|
</video> |
|
|
<div class="carousel-caption"> |
|
|
<h5>Mill19 - Building</h5> |
|
|
</div> |
|
|
</div> |
|
|
<div class="carousel-item"> |
|
|
<video class="w-100" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/rubble_dogaussian_pcl.mp4" type="video/mp4"> |
|
|
</video> |
|
|
<div class="carousel-caption"> |
|
|
<h5>Mill-19 - Rubble</h5> |
|
|
</div> |
|
|
</div> |
|
|
<div class="carousel-item"> |
|
|
<video class="w-100" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/campus_dogaussian_pcl.mp4" type="video/mp4"> |
|
|
</video> |
|
|
<div class="carousel-caption"> |
|
|
<h5>UrbanScene3D - Campus</h5> |
|
|
</div> |
|
|
</div> |
|
|
<div class="carousel-item"> |
|
|
<video class="w-100" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/sciart_dogaussian_pcl.mp4" type="video/mp4"> |
|
|
</video> |
|
|
<div class="carousel-caption"> |
|
|
<h5>UrbanScene3D - Sci-Art</h5> |
|
|
</div> |
|
|
</div> |
|
|
<div class="carousel-item"> |
|
|
<video class="w-100" autoplay="" loop="" muted=""> |
|
|
<source src="assets/videos/residence_dogaussian_pcl.mp4" type="video/mp4"> |
|
|
</video> |
|
|
<div class="carousel-caption"> |
|
|
<h5>UrbanScene3D - Residence</h5> |
|
|
</div> |
|
|
</div> |
|
|
</div> |
|
|
<button class="carousel-control-prev" type="button" data-bs-target="#DoGaussianPclResults" data-bs-slide="prev"> |
|
|
<span class="carousel-control-prev-icon" aria-hidden="true"></span> |
|
|
<span class="visually-hidden">Previous</span> |
|
|
</button> |
|
|
<button class="carousel-control-next" type="button" data-bs-target="#DoGaussianPclResults" data-bs-slide="next"> |
|
|
<span class="carousel-control-next-icon" aria-hidden="true"></span> |
|
|
<span class="visually-hidden">Next</span> |
|
|
</button> |
|
|
</div> |
|
|
|
|
|
<h3 class="mt-5">Training Pipeline of <b>DoGaussian</b></h3> |
|
|
<hr class="hr"> |
|
|
|
|
|
<img class="w-100 mb-3" src="assets/imgs/dogaussian_pipeline.png" alt=""> |
|
|
<p> |
|
|
(1) We first split the scene into K blocks with similar sizes. Each block is extended to a larger size |
|
|
to construct overlapping parts. |
|
|
<br> |
|
|
(2) Subsequently, we assign views and points into different blocks. The shared local 3D Gaussians |
|
|
(connected by solid lines in the figure) are a copy the the global 3D Gaussians. |
|
|
<br> |
|
|
(3) The local 3D Gaussians are then collected and averaged to the global 3D Gaussians in each consensus |
|
|
step, and the global 3D Gaussians are shared with each block before training all blocks. |
|
|
<br> |
|
|
(4) Finally, we use the final global 3D Gaussians to synthesize novel views. |
|
|
</p> |
|
|
|
|
|
<h3 class="mt-5"><b>DoGaussian</b> on novel view synthesis</h3> |
|
|
<hr class="hr"> |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<div class="row align-items-start"> |
|
|
<div id="example4" class="bal-container-small"> |
|
|
<div class="bal-after"> |
|
|
<img src="assets/imgs/campus_dogs.png"> |
|
|
<div class="bal-afterPosition afterLabel" style="z-index:1;"> |
|
|
Ours |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="bal-before" style="width:6.369426751592357%;"> |
|
|
<div class="bal-before-inset" style="width: 628px;"> |
|
|
<img src="assets/imgs/campus_mega-nerf.png"> |
|
|
<div class="bal-beforePosition beforeLabel"> |
|
|
Mega-NeRF [Turki, <i>et.al.</i> 2022] |
|
|
</div> |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="bal-handle" style="left:6.369426751592357%;"> |
|
|
<span class=" handle-left-arrow"></span> |
|
|
<span class="handle-right-arrow"></span> |
|
|
</div> |
|
|
</div> |
|
|
<hr> |
|
|
|
|
|
<div id="example3" class="bal-container-small"> |
|
|
<div class="bal-after"> |
|
|
<img src="assets/imgs/building_dogs.png"> |
|
|
<div class="bal-afterPosition afterLabel" style="z-index:1;"> |
|
|
Ours |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="bal-before" style="width:6.369426751592357%;"> |
|
|
<div class="bal-before-inset" style="width: 628px;"> |
|
|
<img src="assets/imgs/building_switch-nerf.png"> |
|
|
<div class="bal-beforePosition beforeLabel"> |
|
|
Switch-NeRF [Mi & Xu 2023] |
|
|
</div> |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="bal-handle" style="left:6.369426751592357%;"> |
|
|
<span class=" handle-left-arrow"></span> |
|
|
<span class="handle-right-arrow"></span> |
|
|
</div> |
|
|
</div> |
|
|
<hr> |
|
|
</div> |
|
|
|
|
|
<div class="row align-items-start"> |
|
|
<div id="example1" class="bal-container-small"> |
|
|
<div class="bal-after"> |
|
|
<img src="assets/imgs/rubble_dogs.png"> |
|
|
<div class="bal-afterPosition afterLabel" style="z-index:1;"> |
|
|
Ours |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="bal-before" style="width:6.369426751592357%;"> |
|
|
<div class="bal-before-inset" style="width: 628px;"> |
|
|
<img src="assets/imgs/rubble_3dgs.png"> |
|
|
<div class="bal-beforePosition beforeLabel"> |
|
|
3DGS [Kerbl, <i>et.al.</i> 2023] |
|
|
</div> |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="bal-handle" style="left:6.369426751592357%;"> |
|
|
<span class=" handle-left-arrow"></span> |
|
|
<span class="handle-right-arrow"></span> |
|
|
</div> |
|
|
</div> |
|
|
<hr> |
|
|
|
|
|
<div id="example2" class="bal-container-small"> |
|
|
<div class="bal-after"> |
|
|
<img src="assets/imgs/sciart_dogs.png"> |
|
|
<div class="bal-afterPosition afterLabel" style="z-index:1;"> |
|
|
Ours |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="bal-before" style="width:6.369426751592357%;"> |
|
|
<div class="bal-before-inset" style="width: 628px;"> |
|
|
<img src="assets/imgs/sciart_vastgs.png"> |
|
|
<div class="bal-beforePosition beforeLabel"> |
|
|
VastGaussian [Lin, <i>et.al.</i> 2024] |
|
|
</div> |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="bal-handle" style="left:6.369426751592357%;"> |
|
|
<span class=" handle-left-arrow"></span> |
|
|
<span class="handle-right-arrow"></span> |
|
|
</div> |
|
|
</div> |
|
|
<hr> |
|
|
</div> |
|
|
|
|
|
<h3 class="mt-5" id="citation">Please consider citing our paper</h3> |
|
|
<hr class="hr"> |
|
|
<pre class="w-100 user-select-all font-monospace border border-secondary bg-dark mb-5 p-2">@inproceedings{yuchen2024dogaussian, |
|
|
title={DoGaussian: Distributed-Oriented Gaussian Splatting for Large-Scale 3D Reconstruction Via Gaussian Consensus}, |
|
|
author={Yu Chen, Gim Hee Lee}, |
|
|
booktitle={arXiv}, |
|
|
year={2024}, |
|
|
}</pre> |
|
|
|
|
|
</div> |
|
|
<div class="col-sm-0 col-md-2"></div> |
|
|
|
|
|
|
|
|
|
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<script src="assets/bootstrap.bundle.min.js"></script> |
|
|
<script src="https://kit.fontawesome.com/746ee6bfa4.js"></script> |
|
|
|
|
|
<script src="assets/slider.js"></script> |
|
|
<script> |
|
|
new BeforeAfter({ |
|
|
id: '#example1' |
|
|
}); |
|
|
new BeforeAfter({ |
|
|
id: '#example2' |
|
|
}); |
|
|
new BeforeAfter({ |
|
|
id: '#example3' |
|
|
}); |
|
|
new BeforeAfter({ |
|
|
id: '#example4' |
|
|
}); |
|
|
new BeforeAfter({ |
|
|
id: '#example6' |
|
|
}); |
|
|
new BeforeAfter({ |
|
|
id: '#example7' |
|
|
}); |
|
|
new BeforeAfter({ |
|
|
id: '#example8' |
|
|
}); |
|
|
new BeforeAfter({ |
|
|
id: '#example9' |
|
|
}); |
|
|
|
|
|
</script> |
|
|
|
|
|
<script> |
|
|
window.addEventListener('load', videoScroll); |
|
|
window.addEventListener('scroll', videoScroll); |
|
|
|
|
|
function videoScroll() { |
|
|
|
|
|
if ( document.querySelectorAll('video[autoplay]').length > 0) { |
|
|
|
|
|
var windowHeight = window.innerHeight, |
|
|
videoEl = document.querySelectorAll('video[autoplay]'); |
|
|
|
|
|
for (var i = 0; i < videoEl.length; i++) { |
|
|
|
|
|
var thisVideoEl = videoEl[i], |
|
|
videoHeight = thisVideoEl.clientHeight, |
|
|
videoClientRect = thisVideoEl.getBoundingClientRect().top; |
|
|
|
|
|
if ( (thisVideoEl.parentNode.classList.contains('carousel-item') && thisVideoEl.parentNode.classList.contains('active')) || (thisVideoEl.parentNode.nodeName === 'DIV' && !thisVideoEl.parentNode.classList.contains('carousel-item') )) { |
|
|
if ( videoClientRect <= ( (windowHeight) - (videoHeight*.8) ) && videoClientRect >= ( 0 - ( videoHeight*.2 ) ) ) { |
|
|
thisVideoEl.play(); |
|
|
} else { |
|
|
thisVideoEl.pause(); |
|
|
} |
|
|
} |
|
|
} |
|
|
} |
|
|
} |
|
|
</script> |
|
|
|
|
|
</body><grammarly-desktop-integration data-grammarly-shadow-root="true"><template shadowrootmode="open"><style> |
|
|
div.grammarly-desktop-integration { |
|
|
position: absolute; |
|
|
width: 1px; |
|
|
height: 1px; |
|
|
padding: 0; |
|
|
margin: -1px; |
|
|
overflow: hidden; |
|
|
clip: rect(0, 0, 0, 0); |
|
|
white-space: nowrap; |
|
|
border: 0; |
|
|
-moz-user-select: none; |
|
|
-webkit-user-select: none; |
|
|
-ms-user-select:none; |
|
|
user-select:none; |
|
|
} |
|
|
|
|
|
div.grammarly-desktop-integration:before { |
|
|
content: attr(data-content); |
|
|
} |
|
|
</style><div aria-label="grammarly-integration" role="group" tabindex="-1" class="grammarly-desktop-integration" data-content="{"mode":"full","isActive":true,"isUserDisabled":false}"></div></template></grammarly-desktop-integration></html> |
|
|
|
